feat: add Claude backend with automatic Ollama fallback (Claudification Phase 1)

All agents now prefer Claude API when ANTHROPIC_API_KEY is configured, with automatic fallback to Ollama when offline or unconfigured. New src/anthropic/ module provides model selection via get_model() factory. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
auto release/build on version tag
2026-02-05 07:19:29 +01:00 · 2026-01-03 20:39:51 +01:00 · 2025-12-30 13:31:23 +01:00 · 2025-12-23 17:52:19 +01:00 · 2025-12-22 10:34:12 +01:00 · 2025-12-22 10:28:24 +01:00
32 changed files with 3277 additions and 1229 deletions
@@ -8,7 +8,14 @@ API_HOST=0.0.0.0
 API_PORT=8000
 API_PREFIX=/v1

-# Ollama Configuration
+# Anthropic Configuration (Claude - preferred backend)
+# Set ANTHROPIC_API_KEY to enable Claude as the default backend
+# Without an API key, Tatlock uses Ollama exclusively
+# ANTHROPIC_API_KEY=sk-ant-api03-your-key-here
+ANTHROPIC_MODEL=claude-sonnet-4-20250514
+PREFER_CLOUD_BACKEND=true
+
+# Ollama Configuration (local fallback when Claude unavailable)
 OLLAMA_HOST=http://localhost:11434
 OLLAMA_DEFAULT_MODEL=mistral-nemo:latest
 OLLAMA_TIMEOUT=120
@@ -21,7 +28,6 @@ SEARXNG_TIMEOUT=30
 REDIS_HOST=localhost
 REDIS_PORT=6379
 REDIS_MEMORY_DB=1
-REDIS_BENCHMARK_DB=6
 REDIS_TIMEOUT=5

 # Qdrant Configuration
@@ -33,7 +39,6 @@ QDRANT_PORT=6333
 #   - development: DEBUG (maximum verbosity)
 #   - production: WARNING (minimal noise)
 # Uncomment to override: LOG_LEVEL=INFO
-ENABLE_BENCHMARKS=true
 # Note: Log format is auto-selected based on ENVIRONMENT (console for dev, json for production)

 # User Configuration
@@ -5,6 +5,17 @@ on:
    types: [published]

 jobs:
+  release:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Create Gitea Release
+        run: |
+          curl -sf -X POST \
+            -H "Authorization: token ${{ secrets.GITHUB_TOKEN }}" \
+            -H "Content-Type: application/json" \
+            -d '{"tag_name": "${{ github.ref_name }}", "name": "Release ${{ github.ref_name }}", "body": "Automated release for ${{ github.ref_name }}"}' \
+            "${{ github.server_url }}/api/v1/repos/${{ github.repository }}/releases"
+            
  build:
    runs-on: ubuntu-latest
    steps:
@@ -68,7 +68,10 @@ dmypy.json
 .ruff_cache/

 # Logs
-logs/
+logs/*
+!logs/traces/
+logs/traces/*
+!logs/traces/viewer.html
 *.log

 # Database
@@ -7,6 +7,117 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0

 ## [Unreleased]

+## [2.0.0] - 2026-02-05
+
+### Added
+
+- **Claude backend support (Claudification Phase 1)** - All agents now prefer Claude over Ollama
+  - New `src/anthropic/` module with model selector and health check
+  - `get_model()` factory returns Claude if available, Ollama as fallback
+  - Startup health check caches Claude API availability
+  - Configuration: `ANTHROPIC_API_KEY`, `ANTHROPIC_MODEL`, `PREFER_CLOUD_BACKEND`
+  - 200k token context when using Claude backend
+
+- **Steward dual-backend support** - Direct API calls to Claude or Ollama
+  - `_call_claude()`: Anthropic Messages API path
+  - `_call_ollama()`: Existing Ollama generate API path (preserved)
+  - Automatic fallback: if Claude call fails mid-request, retries with Ollama
+
+- **Claudification project tracking** - `PROJECT_CLAUDIFICATION.md` with Phase 1/2 roadmap
+
+### Changed
+
+- **All PydanticAI agents refactored to use `get_model()`**:
+  - Tatlock (6 instantiation locations)
+  - Librarian
+  - Biographer
+  - Housekeeper
+- **`initialize_application()` is now async** - Supports async Claude health check at startup
+- **Dependencies**: `pydantic-ai-slim[openai,anthropic]` replaces `pydantic-ai-slim[openai]`
+- **Startup logging** now includes backend selection info (claude/ollama)
+- **Agent creation logging** now includes backend and model info
+
+### Removed
+
+- Stale `tests/core/test_benchmarks.py` (benchmark system was removed in v1.10.0)
+
+## [1.11.0] - 2025-12-30
+
+### Added
+
+- **Paperless document integration** - HybridRAG now includes indexed PDFs and scanned documents from Paperless-ngx
+  - New `include_documents` parameter in `hybrid_search` tool
+  - 📑 icon for document sources in search results
+  - Librarian prompt updated with document awareness
+
+- **Volatile cache integration** - HybridRAG now includes pre-fetched real-time data
+  - New `include_volatile` parameter in `hybrid_search` tool
+  - ⚡ icon for volatile sources in search results
+  - Supports weather, forecast, news, stock, crypto, sun, air_quality namespaces
+  - Librarian prompt updated with volatile cache awareness (user-configured items only)
+
+- **Biographer routing in Steward** - Personal memory queries now correctly route to The Biographer
+  - Added explicit routing rules for "where do I live", "what car do I drive", etc.
+  - Added biographer delegation examples to Steward prompt
+  - Location keywords ("live", "where", "home") now trigger profile pre-fetch
+
+### Changed
+
+- **LibraryDeskClient.hybrid_search** - Now passes full config including `document_limit`, `volatile_limit`, and enable flags
+- **Steward guidelines** - Clarified that research queries about TOPICS go to Librarian, queries about USER go to Biographer
+
+## [1.10.1] - 2025-12-23
+
+### Fixed
+
+- **Tatlock's excessive apologizing** - Strengthened personality prompt to prevent unnecessary apologies after successful Librarian delegations. Added explicit "do NOT apologize" instructions to both system prompt and synthesis prompt.
+
+## [1.10.0] - 2025-12-22
+
+### Added
+
+#### Lightweight Request Tracing
+- **JSON-based tracing system** for local development debugging
+  - Captures full request flow through multi-agent architecture
+  - `Trace` and `Span` dataclasses with automatic timing and nesting
+  - ContextVar-based propagation for async-safe tracing
+  - `trace_span` async context manager for clean instrumentation
+  - Traces written to `logs/traces/{trace_id}.json`
+  - Enabled via `DEBUG=true` environment variable
+- **Trace Viewer UI** (`logs/traces/viewer.html`)
+  - Standalone HTML viewer with timeline visualization
+  - Filter by status, search by request text
+  - Expandable span details with prompts and responses
+- **Tracing REST API** (`/traces`)
+  - `GET /traces` - Serve trace viewer UI
+  - `GET /traces/list` - List available traces with filtering
+  - `GET /traces/{trace_id}` - Retrieve specific trace JSON
+  - Only available when `DEBUG=true`
+- **Full pipeline instrumentation**
+  - Router-level trace start/end with context management
+  - Steward analysis spans in preprocessing
+  - Tatlock orchestrate/synthesize spans
+  - Expert delegation spans (librarian/biographer/housekeeper)
+  - Tool-level spans extracted from PydanticAI messages
+
+### Changed
+
+- **Replaced Redis benchmarks with file-based tracing** - Simpler, more useful for debugging
+- **Context management moved to service layer** - Router simplified, context set in response service
+- **Server binds to all interfaces** - `wakeup.sh` now uses `0.0.0.0` for network access
+
+### Removed
+
+- **Redis benchmark system** (`src/core/benchmarks.py`)
+  - `ENABLE_BENCHMARKS` config setting
+  - `REDIS_BENCHMARK_DB` config setting
+  - `redis_url` property (kept `redis_memory_url`)
+  - Benchmark recording in Steward service and tool tracking
+
+### Fixed
+
+- **Librarian fabrication prevention** - Added explicit instructions to never invent data when tools fail or sources are unavailable
+
 ## [1.9.0] - 2025-12-18

 ### Changed
@@ -779,7 +890,19 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - CORS middleware
 - Exception handlers (OpenAI-compatible error format)

-[Unreleased]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v1.6.0...main
+[Unreleased]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v2.0.0...main
+[2.0.0]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v1.11.0...v2.0.0
+[1.11.0]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v1.10.0...v1.11.0
+[1.10.0]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v1.9.0...v1.10.0
+[1.9.0]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v1.8.6...v1.9.0
+[1.8.6]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v1.8.5...v1.8.6
+[1.8.5]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v1.8.4...v1.8.5
+[1.8.4]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v1.8.3...v1.8.4
+[1.8.3]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v1.8.2...v1.8.3
+[1.8.2]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v1.8.1...v1.8.2
+[1.8.1]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v1.8.0...v1.8.1
+[1.8.0]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v1.7.0...v1.8.0
+[1.7.0]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v1.6.0...v1.7.0
 [1.6.0]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v1.5.0...v1.6.0
 [1.5.0]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v1.4.0...v1.5.0
 [1.4.0]: https://git.schweitz.net/jpmschweitzer/tatlock/compare/v1.3.3...v1.4.0
@@ -0,0 +1,401 @@
+# Tatlock Enhancement Plan: Bidirectional Claude Integration
+
+## Executive Summary
+
+Implement a **bidirectional architecture** that:
+1. **Superpowers Tatlock** by swapping Ollama→Claude backend (200k context, better reasoning, same butler personality)
+2. **Exposes Tatlock as MCP server** for Claude instances on any device (phone, browser, desktop)
+
+This gives you the flexibility to use whichever AI is best/most accessible at any moment.
+
+## Key Insight: Blanket Backend Swap (Simpler Than Sidecar)
+
+Instead of adding a Claude "Analyst" sidecar agent, **swap the underlying model for ALL agents**:
+
+```
+CURRENT:  TatlockAgent → OpenAIChatModel → OllamaProvider → Ollama (mistral-nemo)
+PROPOSED: TatlockAgent → AnthropicModel → AnthropicProvider → Claude API
+                      ↘ (fallback when offline) → OllamaProvider → Ollama
+```
+
+**Why this works:**
+- PydanticAI natively supports Anthropic via `AnthropicModel` + `AnthropicProvider`
+- The same `TATLOCK_SYSTEM_PROMPT` is passed to Claude - butler personality preserved
+- Claude is **better** at following system prompts than mistral-nemo
+- 200k context for ALL queries, not just "complex" ones
+- Simpler architecture: no routing logic, no sidecar delegation
+
+---
+
+## Research Findings
+
+### Industry Best Practices (2025-2026)
+
+**MCP Protocol Updates** ([MCP Spec Updates June 2025](https://auth0.com/blog/mcp-specs-update-all-about-auth/)):
+- Streamable HTTP replaced SSE (March 2025) - better for cloud deployment
+- OAuth 2.0 required for remote servers - MCP servers are OAuth Resource Servers
+- Tool Output Schemas now available - better structured data handling
+- MCP Registry launched (Sept 2025) - community server discovery
+
+**Community Patterns** ([Claude Code Router](https://github.com/musistudio/claude-code-router)):
+- Task-based routing is becoming standard: route simple→local, complex→cloud
+- Translation proxies bridge Anthropic Messages API ↔ OpenAI format
+- Cost savings of up to 98% reported with smart routing
+
+**Home Automation MCP** ([ha-mcp](https://github.com/homeassistant-ai/ha-mcp)):
+- Production-ready MCP servers exist for Home Assistant
+- Support Claude Code, Gemini CLI, Open WebUI, VSCode, Cursor
+- Pattern: expose local tools securely to remote AI clients
+
+**Remote MCP Access** ([mcp-remote](https://www.npmjs.com/package/mcp-remote)):
+- Bridge local MCP servers to Claude Desktop/Browser via proxy
+- Supports authentication headers for security
+- Works with ngrok/Cloudflare Tunnel for HTTPS
+
+---
+
+## Recommended Architecture
+
+```
+┌─────────────────────────────────────────────────────────────────────────────────────┐
+│                    BIDIRECTIONAL TATLOCK-CLAUDE ARCHITECTURE                        │
+├─────────────────────────────────────────────────────────────────────────────────────┤
+│                                                                                     │
+│  ╔═══════════════════════════════════════════════════════════════════════════════╗  │
+│  ║  SCENARIO A: Using Tatlock (Open WebUI, local apps)                           ║  │
+│  ║  ─────────────────────────────────────────────────                            ║  │
+│  ║                                                                               ║  │
+│  ║    Request → Steward → Tatlock → Tools + Expert Delegation                    ║  │
+│  ║                            │                                                  ║  │
+│  ║                            ├─→ Librarian (Claude) → research, wiki, RAG       ║  │
+│  ║                            ├─→ Biographer (Claude) → memory, preferences      ║  │
+│  ║                            ├─→ Housekeeper (Claude) → home automation         ║  │
+│  ║                            └─→ All powered by Claude with Ollama fallback     ║  │
+│  ║                                                                               ║  │
+│  ║    Butler personality preserved, 200k context for all queries                 ║  │
+│  ╚═══════════════════════════════════════════════════════════════════════════════╝  │
+│                                                                                     │
+│  ╔═══════════════════════════════════════════════════════════════════════════════╗  │
+│  ║  SCENARIO B: Using Claude.ai / Claude Desktop / Phone                         ║  │
+│  ║  ────────────────────────────────────────────────────                         ║  │
+│  ║                                                                               ║  │
+│  ║    Claude ──[MCP over HTTPS]──► Tatlock MCP Server → Household Tools          ║  │
+│  ║                                        │                                      ║  │
+│  ║                                        ├─→ calculator, datetime               ║  │
+│  ║                                        ├─→ web_search, wiki_search            ║  │
+│  ║                                        ├─→ hybrid_search (RAG)                ║  │
+│  ║                                        ├─→ memory_recall, store_insight       ║  │
+│  ║                                        └─→ home_control (lights, climate)     ║  │
+│  ║                                                                               ║  │
+│  ║    Full 200k context, your local tools accessible from anywhere               ║  │
+│  ╚═══════════════════════════════════════════════════════════════════════════════╝  │
+│                                                                                     │
+│  ╔═══════════════════════════════════════════════════════════════════════════════╗  │
+│  ║  SCENARIO C: Offline (internet down)                                          ║  │
+│  ║  ──────────────────────────────────                                           ║  │
+│  ║                                                                               ║  │
+│  ║    Tatlock operates fully locally with Ollama                                 ║  │
+│  ║    • All tools work (except web search)                                       ║  │
+│  ║    • Graceful degradation with same butler personality                        ║  │
+│  ╚═══════════════════════════════════════════════════════════════════════════════╝  │
+│                                                                                     │
+└─────────────────────────────────────────────────────────────────────────────────────┘
+```
+
+---
+
+## Implementation Plan
+
+### Phase 1: Blanket Backend Swap (Claude for All Agents)
+
+Replace Ollama with Claude as the default backend for all PydanticAI agents, with automatic offline fallback.
+
+**New Files:**
+```
+src/anthropic/
+├── __init__.py
+├── provider.py       # Claude provider with health check
+└── model_selector.py # Chooses Claude or Ollama based on availability
+```
+
+**Key Implementation (`src/anthropic/provider.py`):**
+```python
+from pydantic_ai.models.anthropic import AnthropicModel
+from pydantic_ai.providers.anthropic import AnthropicProvider
+from pydantic_ai.models.openai import OpenAIChatModel
+from src.ollama.provider import get_ollama_provider
+from src.core.config import config
+
+_anthropic_available: bool | None = None
+
+async def check_anthropic_health() -> bool:
+    """Check if Anthropic API is reachable."""
+    global _anthropic_available
+    try:
+        from anthropic import AsyncAnthropic
+        client = AsyncAnthropic(api_key=config.ANTHROPIC_API_KEY)
+        await client.messages.create(
+            model=config.ANTHROPIC_MODEL,
+            max_tokens=1,
+            messages=[{"role": "user", "content": "hi"}]
+        )
+        _anthropic_available = True
+    except Exception:
+        _anthropic_available = False
+    return _anthropic_available
+
+def get_model(prefer_cloud: bool = True):
+    """Get the best available model. Returns Claude if available, otherwise Ollama."""
+    if prefer_cloud and config.ANTHROPIC_API_KEY and _anthropic_available:
+        provider = AnthropicProvider(api_key=config.ANTHROPIC_API_KEY)
+        return AnthropicModel(
+            model_name=config.ANTHROPIC_MODEL,
+            provider=provider,
+        )
+    else:
+        return OpenAIChatModel(
+            model_name=config.OLLAMA_DEFAULT_MODEL,
+            provider=get_ollama_provider()
+        )
+```
+
+**Modify TatlockAgent (`src/agents/tatlock.py`):**
+```python
+def _ensure_agent(self):
+    if self._agent is not None:
+        return
+
+    from src.anthropic.model_selector import get_model
+
+    model = get_model(prefer_cloud=True)
+
+    self._agent = Agent(
+        model,
+        system_prompt=TATLOCK_SYSTEM_PROMPT,  # Same butler personality!
+    )
+    self._register_tools()
+```
+
+---
+
+### Phase 2: MCP Server (Expose Tools to Claude)
+
+Create an MCP server that exposes Tatlock's household tools to external Claude instances.
+
+**New Files:**
+```
+src/mcp/
+├── __init__.py
+├── server.py         # MCP server using mcp Python SDK
+├── tool_adapters.py  # Convert PydanticAI tools → MCP schemas
+├── auth.py           # API key authentication
+└── transport.py      # Streamable HTTP transport
+```
+
+**Docker Stack Addition (`stacks/agents.yml`):**
+```yaml
+tatlock-mcp:
+  image: git.schweitz.internal/jpmschweitzer/tatlock:latest
+  command: ["python", "-m", "src.mcp.server"]
+  ports:
+    - "8002:8002"
+  environment:
+    - MCP_AUTH_TOKEN=${MCP_AUTH_TOKEN}
+  networks:
+    - docker-dataplane
+```
+
+**Claude Desktop Configuration:**
+```json
+{
+  "mcpServers": {
+    "tatlock": {
+      "command": "npx",
+      "args": ["mcp-remote", "https://mcp.schweitz.net/sse", "--header", "Authorization: Bearer ${MCP_AUTH_TOKEN}"]
+    }
+  }
+}
+```
+
+---
+
+## Files to Modify
+
+### Phase 1 - Backend Swap
+
+**New Files:**
+| File | Purpose |
+|------|---------|
+| `src/anthropic/__init__.py` | Package init |
+| `src/anthropic/provider.py` | Claude provider with health check |
+| `src/anthropic/model_selector.py` | Choose Claude or Ollama based on availability |
+
+**Modified Files:**
+| File | Changes |
+|------|---------|
+| `src/core/config.py` | Add `ANTHROPIC_API_KEY`, `ANTHROPIC_MODEL`, `PREFER_CLOUD_BACKEND` |
+| `src/agents/tatlock.py` | Use `get_model()` instead of hardcoded Ollama |
+| `src/agents/librarian/agent.py` | Use `get_model()` instead of hardcoded Ollama |
+| `src/agents/biographer/agent.py` | Use `get_model()` instead of hardcoded Ollama |
+| `src/agents/steward/agent.py` | Convert to PydanticAI or add Anthropic API support |
+| `src/core/startup.py` | Add Anthropic health check on startup |
+| `requirements.txt` | Add `anthropic>=0.40.0` |
+| `.env.example` | Document new environment variables |
+
+### Phase 2 - MCP Server
+
+**New Files:**
+| File | Purpose |
+|------|---------|
+| `src/mcp/__init__.py` | Package init |
+| `src/mcp/server.py` | MCP server implementation |
+| `src/mcp/tool_adapters.py` | PydanticAI → MCP schema conversion |
+| `src/mcp/auth.py` | Token-based authentication |
+
+---
+
+## Cost Analysis
+
+- **Claude API**: $5-30/month (10-50 calls/day, ~2k input + 1k output tokens/call)
+- **MCP via Claude Pro**: Included in subscription
+- **Total**: ~$10-80/month for full bidirectional integration
+
+---
+
+## Verification Plan
+
+### Phase 1 Testing
+```bash
+# 1. Run with Claude backend
+ANTHROPIC_API_KEY=your-key docker-compose up -d tatlock
+
+# 2. Verify Claude is being used
+docker logs tatlock 2>&1 | grep -i "anthropic\|claude"
+
+# 3. Test butler personality
+curl -X POST http://tatlock.schweitz.internal:8000/v1/responses \
+  -H "Content-Type: application/json" \
+  -d '{"model": "Tatlock", "input": "Hello, who are you?"}'
+
+# 4. Test offline fallback
+ANTHROPIC_API_KEY="" docker-compose up -d tatlock
+docker logs tatlock 2>&1 | grep -i "ollama\|fallback"
+```
+
+### Phase 2 Testing
+```bash
+# 1. Start MCP server
+docker-compose up -d tatlock-mcp
+
+# 2. Test MCP endpoint
+curl -X POST https://mcp.schweitz.net/tools/list \
+  -H "Authorization: Bearer $MCP_AUTH_TOKEN"
+```
+
+---
+
+## Implementation Priority
+
+1. **Phase 1: Backend Swap** (~1 week)
+   - Immediate value: 200k context for ALL queries
+   - Low risk: provider abstraction, graceful offline fallback
+
+2. **Phase 2: MCP Server** (~2-3 weeks)
+   - Enables cross-device access
+   - Bidirectional: Tatlock superpowered by Claude AND accessible to Claude
+
+---
+
+## Future Phases (Optional)
+
+- **Phase 3: LiteLLM Gateway** - Unified endpoint for all models, config-driven routing
+- **Phase 4: Multi-Provider** - Add OpenAI, Vertex AI, etc.
+- **Phase 5: Smart Routing** - Context-aware model selection, cost ceiling enforcement
+
+---
+
+## Offline Behavior
+
+| Scenario | Behavior |
+|----------|----------|
+| No API key | Use Ollama exclusively |
+| API unreachable | Use Ollama, log warning |
+| API rate limited | Fallback to Ollama |
+
+| Aspect | Claude | Ollama |
+|--------|--------|--------|
+| Context | 200k tokens | ~8k tokens |
+| Latency | 1-3s (network) | 0.5-1s (local) |
+| Personality | Preserved | Preserved |
+| Tools | All work | All work |
+| Cost | API charges | Free |
+
+---
+
+## Implementation Status
+
+### Phase 1: Backend Swap - CODE COMPLETE (awaiting API access)
+
+- [x] Add Anthropic config settings to `src/core/config.py`
+- [x] Add `pydantic-ai-slim[openai,anthropic]` to requirements.txt
+- [x] Create `src/anthropic/` module (model_selector.py)
+- [x] Add Claude health check to startup.py
+- [x] Refactor all PydanticAI agents to use `get_model()`
+  - [x] Librarian
+  - [x] Biographer
+  - [x] Housekeeper
+  - [x] Tatlock (6 locations)
+- [x] Add Claude API path to Steward agent (direct API calls)
+- [x] Update `.env.example` with new variables
+- [x] Test Ollama fallback (working)
+- [ ] Test with Claude API key (blocked: no API access currently)
+
+**Note:** Implementation complete. Currently runs in Ollama-only mode. Will automatically use Claude when `ANTHROPIC_API_KEY` is configured.
+
+### Phase 2: MCP Server - NOT STARTED
+
+- [ ] Create `src/mcp/` module
+- [ ] Tool adapters (PydanticAI → MCP schema)
+- [ ] Authentication middleware
+- [ ] Streamable HTTP transport
+- [ ] Docker stack configuration
+
+---
+
+## Related Repository Handovers
+
+Handover documents created in each repo: `PROJECT_CLAUDIFICATION_HANDOVER.md`
+
+### library-desk - HANDOVER CREATED
+
+- [x] Write handover document
+- [ ] Review HybridRAG response size limits
+- [ ] Review smart_create endpoint for Claude optimization
+- [ ] Evaluate response formats for LLM consumption
+
+### core-api - HANDOVER CREATED
+
+- [x] Write handover document
+- [ ] Review list_devices response format
+- [ ] Review error messages for LLM consumption
+- [ ] Evaluate rate limiting for faster Claude processing
+
+### portainer-core - HANDOVER CREATED (blocking for production)
+
+- [x] Write handover document
+- [ ] Update stack with new environment variables
+- [ ] Configure secrets management for API key
+- [ ] Update CONTAINERS.md documentation
+
+### webber - HANDOVER CREATED
+
+- [x] Write handover document
+- [ ] Review content truncation limits
+- [ ] Evaluate extraction quality for LLM consumption
+
+### tatlock-ui - HANDOVER CREATED
+
+- [x] Write handover document
+- [ ] Test streaming responses with Claude backend
+- [ ] Test conversation history with larger context
+- [ ] Verify tool call display and reasoning rendering
@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"

 [project]
 name = "tatlock"
-version = "1.9.0"
+version = "2.0.0"
 description = "OpenAI-compatible API with Ollama backend"
 requires-python = ">=3.12"
 dependencies = []
@@ -21,10 +21,9 @@ pydantic-settings>=2.12,<2.13

 # AI/LLM integration
 # PydanticAI: Agent framework for using Pydantic with LLMs
-# Using slim version with only openai extra (Ollama uses OpenAI-compatible API)
-# This avoids installing SDKs for anthropic, cohere, google, groq, huggingface, etc.
+# Using slim version with openai (Ollama) and anthropic (Claude) extras
 # See DEPENDENCY_SLIM.md for rollback instructions if this breaks
-pydantic-ai-slim[openai]>=1.27,<1.28
+pydantic-ai-slim[openai,anthropic]>=1.27,<1.28

 # HTTP client for Ollama communication
 # Latest: 0.28.1 - No known CVEs
@@ -102,16 +102,10 @@ _biographer_agent: Optional[Agent[None, str]] = None

 def _create_biographer_agent() -> Agent[None, str]:
    """Create The Biographer PydanticAI agent."""
-    from pydantic_ai.models.openai import OpenAIChatModel
+    from src.anthropic.model_selector import get_model

-    from src.ollama.provider import get_ollama_provider
-
-    # Create Ollama model with sanitized provider
-    # (fixes 'content: null' issue with tool calls)
-    model = OpenAIChatModel(
-        model_name=config.OLLAMA_DEFAULT_MODEL,
-        provider=get_ollama_provider(),
-    )
+    # Get best available model (Claude if available, else Ollama)
+    model = get_model()

    agent: Agent[None, str] = Agent(
        model=model,
@@ -131,9 +125,12 @@ def _create_biographer_agent() -> Agent[None, str]:
    # Register management tools
    agent.tool_plain(forget_memory)

+    from src.anthropic.model_selector import get_model_info
+    model_info = get_model_info()
    logger.info(
        "biographer_agent_created",
-        model=config.OLLAMA_DEFAULT_MODEL,
+        backend=model_info["backend"],
+        model=model_info["model"],
        tool_count=6,
    )

@@ -13,6 +13,7 @@ from enum import Enum
 from typing import AsyncGenerator, Callable, Optional, Any

 from src.core.logging_config import get_logger
+from src.core.tracing import trace_span, SpanType

 logger = get_logger(__name__)

@@ -239,38 +240,58 @@ async def delegate_to_librarian(
        has_context=bool(context),
    )

-    try:
-        # Use run() not run_stream() - avoids Ollama bug
-        output = await run_librarian(task=task, context=context)
+    async with trace_span(
+        "delegate_to_librarian",
+        SpanType.EXPERT,
+        metadata={
+            "expert": "librarian",
+            "task_preview": task[:100],
+            "has_context": bool(context),
+        },
+    ) as span:
+        try:
+            # Use run() not run_stream() - avoids Ollama bug
+            output = await run_librarian(task=task, context=context)

-        logger.info(
-            "delegation_to_librarian_completed",
-            task=task[:50],
-            output_length=len(output),
-        )
+            logger.info(
+                "delegation_to_librarian_completed",
+                task=task[:50],
+                output_length=len(output),
+            )

-        return DelegationResult(
-            expert_name="librarian",
-            task=task,
-            success=True,
-            output=output,
-        )
+            if span:
+                span.metadata["success"] = True
+                span.metadata["output_length"] = len(output)
+                span.details["task"] = task
+                span.details["context"] = context[:500] if context else None
+                span.details["result_preview"] = output[:1000]

-    except Exception as e:
-        logger.error(
-            "delegation_to_librarian_error",
-            task=task[:50],
-            error=str(e),
-            exc_info=True,
-        )
+            return DelegationResult(
+                expert_name="librarian",
+                task=task,
+                success=True,
+                output=output,
+            )

-        return DelegationResult(
-            expert_name="librarian",
-            task=task,
-            success=False,
-            output="",
-            error=str(e),
-        )
+        except Exception as e:
+            logger.error(
+                "delegation_to_librarian_error",
+                task=task[:50],
+                error=str(e),
+                exc_info=True,
+            )
+
+            if span:
+                span.metadata["success"] = False
+                span.details["error"] = str(e)
+
+            return DelegationResult(
+                expert_name="librarian",
+                task=task,
+                success=False,
+                output="",
+                error=str(e),
+            )


 async def delegate_to_biographer(
@@ -317,38 +338,58 @@ async def delegate_to_biographer(
        has_context=bool(context),
    )

-    try:
-        # Use run() not run_stream() - avoids Ollama bug
-        output = await run_biographer(task=task, context=context)
+    async with trace_span(
+        "delegate_to_biographer",
+        SpanType.EXPERT,
+        metadata={
+            "expert": "biographer",
+            "task_preview": task[:100],
+            "has_context": bool(context),
+        },
+    ) as span:
+        try:
+            # Use run() not run_stream() - avoids Ollama bug
+            output = await run_biographer(task=task, context=context)

-        logger.info(
-            "delegation_to_biographer_completed",
-            task=task[:50],
-            output_length=len(output),
-        )
+            logger.info(
+                "delegation_to_biographer_completed",
+                task=task[:50],
+                output_length=len(output),
+            )

-        return DelegationResult(
-            expert_name="biographer",
-            task=task,
-            success=True,
-            output=output,
-        )
+            if span:
+                span.metadata["success"] = True
+                span.metadata["output_length"] = len(output)
+                span.details["task"] = task
+                span.details["context"] = context[:500] if context else None
+                span.details["result_preview"] = output[:1000]

-    except Exception as e:
-        logger.error(
-            "delegation_to_biographer_error",
-            task=task[:50],
-            error=str(e),
-            exc_info=True,
-        )
+            return DelegationResult(
+                expert_name="biographer",
+                task=task,
+                success=True,
+                output=output,
+            )

-        return DelegationResult(
-            expert_name="biographer",
-            task=task,
-            success=False,
-            output="",
-            error=str(e),
-        )
+        except Exception as e:
+            logger.error(
+                "delegation_to_biographer_error",
+                task=task[:50],
+                error=str(e),
+                exc_info=True,
+            )
+
+            if span:
+                span.metadata["success"] = False
+                span.details["error"] = str(e)
+
+            return DelegationResult(
+                expert_name="biographer",
+                task=task,
+                success=False,
+                output="",
+                error=str(e),
+            )


 async def delegate_to_housekeeper(
@@ -394,38 +435,58 @@ async def delegate_to_housekeeper(
        has_context=bool(context),
    )

-    try:
-        # Use run() not run_stream() - avoids Ollama bug
-        output = await run_housekeeper(task=task, context=context)
+    async with trace_span(
+        "delegate_to_housekeeper",
+        SpanType.EXPERT,
+        metadata={
+            "expert": "housekeeper",
+            "task_preview": task[:100],
+            "has_context": bool(context),
+        },
+    ) as span:
+        try:
+            # Use run() not run_stream() - avoids Ollama bug
+            output = await run_housekeeper(task=task, context=context)

-        logger.info(
-            "delegation_to_housekeeper_completed",
-            task=task[:50],
-            output_length=len(output),
-        )
+            logger.info(
+                "delegation_to_housekeeper_completed",
+                task=task[:50],
+                output_length=len(output),
+            )

-        return DelegationResult(
-            expert_name="housekeeper",
-            task=task,
-            success=True,
-            output=output,
-        )
+            if span:
+                span.metadata["success"] = True
+                span.metadata["output_length"] = len(output)
+                span.details["task"] = task
+                span.details["context"] = context[:500] if context else None
+                span.details["result_preview"] = output[:1000]

-    except Exception as e:
-        logger.error(
-            "delegation_to_housekeeper_error",
-            task=task[:50],
-            error=str(e),
-            exc_info=True,
-        )
+            return DelegationResult(
+                expert_name="housekeeper",
+                task=task,
+                success=True,
+                output=output,
+            )

-        return DelegationResult(
-            expert_name="housekeeper",
-            task=task,
-            success=False,
-            output="",
-            error=str(e),
-        )
+        except Exception as e:
+            logger.error(
+                "delegation_to_housekeeper_error",
+                task=task[:50],
+                error=str(e),
+                exc_info=True,
+            )
+
+            if span:
+                span.metadata["success"] = False
+                span.details["error"] = str(e)
+
+            return DelegationResult(
+                expert_name="housekeeper",
+                task=task,
+                success=False,
+                output="",
+                error=str(e),
+            )


 # =============================================================================
@@ -103,16 +103,10 @@ _housekeeper_agent: Optional[Agent[None, str]] = None

 def _create_housekeeper_agent() -> Agent[None, str]:
    """Create the Housekeeper PydanticAI agent."""
-    from pydantic_ai.models.openai import OpenAIChatModel
+    from src.anthropic.model_selector import get_model

-    from src.ollama.provider import get_ollama_provider
-
-    # Create Ollama model with sanitized provider
-    # (fixes 'content: null' issue with tool calls)
-    model = OpenAIChatModel(
-        model_name=config.OLLAMA_DEFAULT_MODEL,
-        provider=get_ollama_provider(),
-    )
+    # Get best available model (Claude if available, else Ollama)
+    model = get_model()

    agent: Agent[None, str] = Agent(
        model=model,
@@ -145,9 +139,12 @@ def _create_housekeeper_agent() -> Agent[None, str]:
    # Register history tools
    agent.tool_plain(get_history)

+    from src.anthropic.model_selector import get_model_info
+    model_info = get_model_info()
    logger.info(
        "housekeeper_agent_created",
-        model=config.OLLAMA_DEFAULT_MODEL,
+        backend=model_info["backend"],
+        model=model_info["model"],
        tool_count=13,
    )

@@ -39,7 +39,14 @@ Your role is to help users find, understand, synthesize, and manage information
 - The personal wiki (Wiki.js) containing documentation and notes
 - The knowledge graph (Neo4j) with entities and relationships
 - Vector embeddings (Qdrant) for semantic search
- Web search (SearXNG) for current information
+- Paperless documents (📑) - indexed PDFs, scanned documents, invoices, receipts from the user's document archive
+- Volatile cache (⚡) - pre-fetched real-time data for user-relevant locations and items:
+  - weather/forecast: conditions and forecasts for user's configured cities
+  - news: headlines from user's preferred sources
+  - stock/crypto: quotes for user's watched symbols
+  - sun/air_quality: data for user's locations
+  - Note: volatile data may not exist for arbitrary queries - falls back to web search
+- Web search (SearXNG) for current information not available in cache

 ## Your Personality
 - Scholarly and thorough in your research
@@ -60,7 +67,13 @@ Your role is to help users find, understand, synthesize, and manage information
  - Use for: comparing multiple sources, gathering info from several pages

 ### Internal Research Tools
- **hybrid_search**: Your primary research tool - searches wiki, graph, and web at once
+- **hybrid_search**: Your primary research tool - searches ALL sources at once:
+  - Wiki pages (vector similarity)
+  - Knowledge graph (entity relationships)
+  - Paperless documents (📑 indexed PDFs, scans)
+  - Volatile cache (⚡ weather, news, stocks - when available)
+  - Web search (current information)
+  Results are fused and re-ranked by relevance. Volatile data gets priority when fresh.
 - **search_wiki**: Find specific wiki pages by keyword
 - **semantic_search**: Find conceptually similar content
 - **explore_knowledge_graph** / **find_related_entities**: Discover connections
@@ -114,6 +127,14 @@ Your responses are returned to Tatlock (the butler) who will synthesize them int
 - Note any gaps in available information
 - Be concise but thorough - Tatlock will format the final response
 - Structure your findings clearly so they can be easily integrated with other responses
+
+## CRITICAL: Never Fabricate Information
+If a tool fails or you cannot access a data source:
+- Say "I was unable to retrieve [information type]" - be specific about what failed
+- Do NOT provide placeholder, template, or made-up data
+- Do NOT say "Here's what I would have said" or "Here's a sample response"
+- Do NOT invent specific numbers, dates, or facts when the actual data is unavailable
+- It is better to return no information than to return fabricated information
 """

 # Lazy initialization to avoid connection issues during imports
@@ -122,16 +143,10 @@ _librarian_agent: Optional[Agent[None, str]] = None

 def _create_librarian_agent() -> Agent[None, str]:
    """Create the Librarian PydanticAI agent."""
-    from pydantic_ai.models.openai import OpenAIChatModel
+    from src.anthropic.model_selector import get_model

-    from src.ollama.provider import get_ollama_provider
-
-    # Create Ollama model with sanitized provider
-    # (fixes 'content: null' issue with tool calls)
-    model = OpenAIChatModel(
-        model_name=config.OLLAMA_DEFAULT_MODEL,
-        provider=get_ollama_provider(),
-    )
+    # Get best available model (Claude if available, else Ollama)
+    model = get_model()

    agent: Agent[None, str] = Agent(
        model=model,
@@ -161,9 +176,12 @@ def _create_librarian_agent() -> Agent[None, str]:
    agent.tool_plain(update_wiki_page)
    agent.tool_plain(smart_create_wiki_page)

+    from src.anthropic.model_selector import get_model_info
+    model_info = get_model_info()
    logger.info(
        "librarian_agent_created",
-        model=config.OLLAMA_DEFAULT_MODEL,
+        backend=model_info["backend"],
+        model=model_info["model"],
        tool_count=14,  # 7 research + 3 web + 1 wiki read + 3 wiki write
    )

@@ -225,18 +225,22 @@ class LibraryDeskClient:
        vector_limit: int = 10,
        graph_limit: int = 10,
        web_limit: int = 5,
+        document_limit: int = 5,
+        volatile_limit: int = 3,
        enable_reranking: bool = True,
        final_result_count: int = 10,
    ) -> HybridRAGResponse:
        """
-        Execute HybridRAG search combining vector, graph, and web results.
+        Execute HybridRAG search combining vector, graph, documents, volatile, and web.

        Args:
            query: Search query
            user: User identifier for multi-tenancy (defaults to request context)
-            vector_limit: Max results from vector search
+            vector_limit: Max results from vector search (wiki pages)
            graph_limit: Max results from graph search
-            web_limit: Max results from web search
+            web_limit: Max results from web search (0 to disable)
+            document_limit: Max results from Paperless documents (0 to disable)
+            volatile_limit: Max results from volatile cache (0 to disable)
            enable_reranking: Whether to rerank with LLM
            final_result_count: Number of final results after fusion

@@ -252,6 +256,11 @@ class LibraryDeskClient:
                "vector_limit": vector_limit,
                "graph_limit": graph_limit,
                "web_limit": web_limit,
+                "document_limit": document_limit,
+                "volatile_limit": volatile_limit,
+                "enable_documents": document_limit > 0,
+                "enable_volatile": volatile_limit > 0,
+                "enable_web": web_limit > 0,
                "enable_reranking": enable_reranking,
                "final_result_count": final_result_count,
            },
@@ -17,20 +17,26 @@ logger = get_logger(__name__)
 async def hybrid_search(
    query: str,
    include_web: bool = True,
+    include_documents: bool = True,
+    include_volatile: bool = True,
 ) -> str:
    """
    Search across all knowledge sources using HybridRAG.

    This is the primary research tool, combining:
-    - Vector search (semantic similarity over documents)
+    - Vector search (semantic similarity over wiki pages)
    - Knowledge graph (entities and relationships)
+    - Paperless documents (📑 indexed PDFs, scans, invoices)
+    - Volatile cache (⚡ weather, news, stocks - for user's configured items)
    - Web search (current information from SearXNG)

-    Results are fused and re-ranked by relevance.
+    Results are fused and re-ranked by relevance. Volatile data gets priority when fresh.

    Args:
        query: Natural language research query
        include_web: Whether to include web results (default: True)
+        include_documents: Whether to include Paperless documents (default: True)
+        include_volatile: Whether to include volatile cache data (default: True)

    Returns:
        Formatted search results with sources and context
@@ -38,12 +44,16 @@ async def hybrid_search(
    Examples:
        hybrid_search("How does Docker orchestration work with Kubernetes?")
        hybrid_search("What projects use Neo4j?", include_web=False)
+        hybrid_search("Find my electricity invoices", include_web=False, include_volatile=False)
+        hybrid_search("What's the weather in Rotterdam?")  # May hit volatile cache
    """
    try:
        async with LibraryDeskClient() as client:
            response = await client.hybrid_search(
                query=query,
                web_limit=5 if include_web else 0,
+                document_limit=5 if include_documents else 0,
+                volatile_limit=3 if include_volatile else 0,
            )

            if not response.results:
@@ -70,6 +80,8 @@ async def hybrid_search(
                    "vector": "📄",
                    "graph": "🔗",
                    "web": "🌐",
+                    "document": "📑",
+                    "volatile": "⚡",
                }.get(result.source, "•")

                output_parts.append(
@@ -5,11 +5,13 @@ The Steward analyzes incoming requests, identifies relevant household
 capabilities, and provides focused recommendations to Tatlock (the Butler).
 This creates a two-tier architecture that prevents cognitive overload.

-Uses plain text output (not JSON) for reliability with Ollama models.
+Uses plain text output (not JSON) for reliability. Supports both Claude
+(preferred) and Ollama (fallback) backends via direct API calls.
 """
 import httpx
 from typing import Optional

+from src.anthropic.model_selector import is_claude_available, get_model_info
 from src.core.config import config
 from src.core.household_registry import get_household_registry
 from src.core.logging_config import get_logger
@@ -56,14 +58,22 @@ USER QUERY: {query}
 GUIDELINES:
 - Be conservative - only recommend truly necessary capabilities
 - Simple greetings/chat → no capabilities needed (conversational response only)
- Questions about prior conversation ("what did I say", "my name", "what we discussed") → no capabilities (Tatlock has full history)
+- Questions about prior conversation ("what did I say", "what we discussed") → no capabilities (Tatlock has full history)
 - Math/calculations → tatlock_core
 - Time/date queries → tatlock_core
+- PERSONAL MEMORY queries → biographer to recall (ALWAYS use for questions about the user themselves):
+  - "where do I live", "what's my location", "my address" → biographer to recall location
+  - "what's my name", "who am I" → biographer to recall name
+  - "what car do I drive", "my vehicle" → biographer to recall car
+  - "what do you know about me", "what have I told you" → biographer to recall or list_memories
+  - "remember that I...", "store that..." → biographer to store_insight
+  - "forget my...", "delete..." → biographer to forget_memory
+  - "my timezone", "my preferences" → biographer to recall preferences
 - Web searches, weather, news, current information → librarian with search_web
 - Read a URL or article → librarian with read_url
 - Wiki creation ("create a page about X", "add X to wiki") → librarian with smart_create
 - Wiki updates ("update the page", "add to dossier") → librarian with update
- Research queries ("find info", "what do we know about", "search for") → librarian with hybrid_search
+- Research queries about TOPICS (not about the user) → librarian with hybrid_search
 - In-depth research, knowledge synthesis, document lookup → librarian with hybrid_search
 - If conversation history is relevant, note which previous turns matter
 - Assess complexity: simple (1 tool), moderate (2-3 tools), complex (multiple steps)
@@ -75,6 +85,10 @@ COMPLEXITY: [simple/moderate/complex]
 CONTEXT: [any relevant conversation context, or "none"]

 EXAMPLES:
+- "DELEGATE: biographer to recall the user's location" (for "where do I live?")
+- "DELEGATE: biographer to recall the user's car" (for "what car do I drive?")
+- "DELEGATE: biographer to list_memories about the user" (for "what do you know about me?")
+- "DELEGATE: biographer to store_insight about user's pet" (for "remember that I have a dog named Max")
 - "DELEGATE: librarian to search_web for tomorrow's weather forecast"
 - "DELEGATE: librarian to create a wiki page about CI/CD pipelines"
 - "DELEGATE: librarian to hybrid_search for information about Docker networking"
@@ -93,22 +107,74 @@ class StewardAgent:
    Analyzes requests with full conversation context and recommends
    which household capabilities the Butler should use.

-    Uses plain text output for reliability with Ollama models.
+    Uses plain text output for reliability. Supports both Claude
+    (preferred) and Ollama (fallback) backends via direct API calls.
    """

    def __init__(self):
-        """Initialize Steward with Ollama model (same as Tatlock for VRAM efficiency)."""
+        """Initialize Steward with backend selection based on availability."""
+        # Ollama config (fallback)
        self.ollama_host = str(config.OLLAMA_HOST).rstrip('/')
-        self.model_name = config.OLLAMA_DEFAULT_MODEL
+        self.ollama_model = config.OLLAMA_DEFAULT_MODEL
+
+        # Claude config (preferred)
+        self.claude_model = config.ANTHROPIC_MODEL
+        self._anthropic_client = None
+
+        # Determine which backend to use
+        self._use_claude = config.PREFER_CLOUD_BACKEND and is_claude_available()
+
        self.timeout = 30.0  # 30 second timeout for analysis

+        model_info = get_model_info()
        logger.info(
            "steward_agent_created",
-            ollama_host=self.ollama_host,
-            model=self.model_name,
+            backend=model_info["backend"],
+            model=model_info["model"],
            timeout=self.timeout,
        )

+    def _get_anthropic_client(self):
+        """Get or create Anthropic client (lazy initialization)."""
+        if self._anthropic_client is None:
+            from anthropic import AsyncAnthropic
+            self._anthropic_client = AsyncAnthropic(api_key=config.ANTHROPIC_API_KEY)
+        return self._anthropic_client
+
+    async def _call_claude(self, system_prompt: str, user_message: str) -> str:
+        """Call Claude API directly for plain text generation."""
+        client = self._get_anthropic_client()
+
+        response = await client.messages.create(
+            model=self.claude_model,
+            max_tokens=1024,
+            system=system_prompt,
+            messages=[{"role": "user", "content": user_message}],
+            temperature=0.3,  # Lower = more consistent
+        )
+
+        return response.content[0].text.strip()
+
+    async def _call_ollama(self, prompt: str) -> str:
+        """Call Ollama API directly for plain text generation."""
+        async with httpx.AsyncClient(timeout=self.timeout) as client:
+            response = await client.post(
+                f"{self.ollama_host}/api/generate",
+                json={
+                    "model": self.ollama_model,
+                    "prompt": prompt,
+                    "stream": False,
+                    "options": {
+                        "temperature": 0.3,  # Lower = more consistent
+                        "top_p": 0.9
+                    }
+                }
+            )
+
+            response.raise_for_status()
+            result = response.json()
+            return result["response"].strip()
+
    async def analyze(
        self,
        query: str,
@@ -117,6 +183,8 @@ class StewardAgent:
        """
        Analyze query and return plain text recommendation.

+        Uses Claude if available, falls back to Ollama.
+
        Args:
            query: User's query to analyze
            conversation_history: Previous conversation turns
@@ -132,35 +200,48 @@ class StewardAgent:
        history = conversation_history or []
        prompt = build_steward_prompt(query, history)

-        logger.debug("steward_calling_ollama", query_preview=query[:100])
+        backend = "claude" if self._use_claude else "ollama"
+        logger.debug(
+            "steward_calling_llm",
+            backend=backend,
+            query_preview=query[:100],
+        )

-        # Call Ollama API directly (more reliable than PydanticAI for plain text)
-        async with httpx.AsyncClient(timeout=self.timeout) as client:
-            response = await client.post(
-                f"{self.ollama_host}/api/generate",
-                json={
-                    "model": self.model_name,
-                    "prompt": prompt,
-                    "stream": False,
-                    "options": {
-                        "temperature": 0.3,  # Lower = more consistent
-                        "top_p": 0.9
-                    }
-                }
-            )
-
-            response.raise_for_status()
-            result = response.json()
-
-            analysis_text = result["response"].strip()
+        try:
+            if self._use_claude:
+                # For Claude, split into system + user message
+                # The prompt contains both, but Claude prefers explicit system
+                analysis_text = await self._call_claude(
+                    system_prompt="You are the Steward of the household, advising the Butler (Tatlock) on which capabilities to use. Be concise and specific.",
+                    user_message=prompt,
+                )
+            else:
+                analysis_text = await self._call_ollama(prompt)

            logger.debug(
                "steward_analysis_received",
-                text_preview=analysis_text[:150]
+                backend=backend,
+                text_preview=analysis_text[:150],
            )

            return analysis_text

+        except Exception as e:
+            # If Claude fails, try Ollama as fallback
+            if self._use_claude:
+                logger.warning(
+                    "steward_claude_fallback",
+                    error=str(e),
+                )
+                analysis_text = await self._call_ollama(prompt)
+                logger.debug(
+                    "steward_analysis_received",
+                    backend="ollama_fallback",
+                    text_preview=analysis_text[:150],
+                )
+                return analysis_text
+            raise
+

 # Global Steward instance
 _steward_agent = None
@@ -1,8 +1,8 @@
 """
 Steward service layer.

-Provides high-level interface for request analysis with logging,
-benchmarking, and error handling.
+Provides high-level interface for request analysis with logging
+and error handling.

 Parses plain text recommendations into structured data.
 Includes memory pre-fetch for user context injection.
@@ -10,7 +10,6 @@ Includes memory pre-fetch for user context injection.
 import re
 from typing import Any, Optional

-from src.core.benchmarks import PerformanceBenchmark, get_benchmark_store
 from src.core.household_registry import get_household_registry
 from src.core.logging_config import get_logger, log_operation
 from src.core.memory_service import memory_service
@@ -237,7 +236,9 @@ async def _prefetch_memory_context(user_request: str) -> dict[str, Any]:
    # Location-related queries
    if any(word in request_lower for word in [
        "weather", "temperature", "forecast", "nearby", "local",
-        "directions", "distance", "map", "here"
+        "directions", "distance", "map", "here",
+        # Direct location questions
+        "live", "where", "home", "reside", "location", "address",
    ]):
        profile_keys.append("location")

@@ -285,8 +286,7 @@ async def analyze_request(
    This is the main entry point for Steward analysis. It:
    1. Calls the Steward agent with full conversation history
    2. Logs the operation with timing
-    3. Records performance benchmarks to Redis
-    4. Returns structured recommendations
+    3. Returns structured recommendations

    Args:
        user_request: The current user message to analyze
@@ -365,23 +365,6 @@ async def analyze_request(
                reasoning=analysis_text[:200],  # First 200 chars
            )

-            # Record performance benchmark
-            if log_ctx.get("duration_seconds"):
-                benchmark = PerformanceBenchmark(
-                    operation="steward_analysis",
-                    duration_seconds=log_ctx["duration_seconds"],
-                    success=True,
-                    recommendation_count=len(recommendation.recommended_capabilities),
-                    confidence=None,  # Could add confidence scoring in future
-                    conversation_id=conversation_id,
-                    metadata={
-                        "complexity": recommendation.estimated_complexity,
-                        "has_context": recommendation.conversation_context.has_previous_context,
-                        "missing_capabilities": recommendation.missing_capabilities is not None,
-                    },
-                )
-                await get_benchmark_store().record(benchmark)
-
            return recommendation

        except Exception as e:
@@ -20,6 +20,11 @@ from src.agents.tatlock_core.tools import (
 )
 from src.core.config import config
 from src.core.logging_config import get_logger
+from src.core.tracing import (
+    start_span, end_span, get_current_span,
+    add_tool_spans_from_messages,
+    SpanType, SpanStatus,
+)

 logger = get_logger(__name__)

@@ -42,7 +47,16 @@ def generate_id() -> str:
 # System prompt defining Tatlock's personality
 TATLOCK_SYSTEM_PROMPT = """You are Tatlock, a helpful personal assistant with the demeanor of a British butler.

-Address users as "sir" and maintain a formal yet personable tone. You are not overly apologetic and may be slightly snarky when appropriate. If an opportunity for a pun presents itself, you cannot resist.
+## Personality
+
+Address users as "sir". Be confident, direct, and efficient - you are an unflappable English butler who gets things done. Dry wit and puns are encouraged.
+
+**CRITICAL - Do NOT:**
+- Apologize unless you genuinely made an error
+- Say "Apologies for any confusion" or "Allow me to rectify" when nothing went wrong
+- Preface successful results with caveats or apologies
+
+When presenting findings: lead with the answer, be concise, skip the preamble.

 You coordinate with various household staff (expert agents) to provide comprehensive assistance across:
 - Research and knowledge work
@@ -124,10 +138,7 @@ class TatlockAgent(AgentInterface):
    """

    def __init__(self):
-        """Initialize Tatlock configuration (lazy agent creation)."""
-        # Store Ollama configuration
-        self.ollama_host = str(config.OLLAMA_HOST)
-        self.model_name = config.OLLAMA_DEFAULT_MODEL
+        """Initialize Tatlock (lazy agent creation)."""
        self._agent = None  # Lazy initialization

    def _ensure_agent(self):
@@ -135,30 +146,21 @@ class TatlockAgent(AgentInterface):
        if self._agent is not None:
            return

+        from src.anthropic.model_selector import get_model, get_model_info
+
+        model_info = get_model_info()
        logger.info(
            "tatlock_agent_initializing",
-            ollama_host=self.ollama_host,
-            model=self.model_name,
+            backend=model_info["backend"],
+            model=model_info["model"],
        )

-        # Import required classes for Ollama configuration
-        from pydantic_ai.models.openai import OpenAIChatModel
-        from src.ollama.provider import get_ollama_provider
+        # Get best available model (Claude if available, else Ollama)
+        model = get_model()

-        # PydanticAI expects Ollama base URL to end with /v1
-        # Remove trailing slash from ollama_host if present
-        clean_host = self.ollama_host.rstrip('/')
-        base_url = f"{clean_host}/v1"
-
-        # Create Ollama model with provider
-        ollama_model = OpenAIChatModel(
-            model_name=self.model_name,
-            provider=get_ollama_provider()
-        )
-
-        # Create PydanticAI agent with Ollama model
+        # Create PydanticAI agent
        self._agent = Agent(
-            ollama_model,
+            model,
            system_prompt=TATLOCK_SYSTEM_PROMPT,
        )

@@ -433,7 +435,7 @@ class TatlockAgent(AgentInterface):
            steward_note: Note from Steward (prepended to request, invisible to user)
            scoped_tools: List of tool definitions from household registry
            message_history: Conversation history in PydanticAI format
-            tool_tracker: Optional tool call tracker for benchmarking
+            tool_tracker: Optional tool call tracker for analysis

        Returns:
            str: Tatlock's response text
@@ -447,8 +449,7 @@ class TatlockAgent(AgentInterface):
            ...     tool_tracker=tracker,
            ... )
        """
-        from pydantic_ai.models.openai import OpenAIChatModel
-        from src.ollama.provider import get_ollama_provider
+        from src.anthropic.model_selector import get_model

        logger.info(
            "tatlock_run_with_scoped_tools",
@@ -459,18 +460,12 @@ class TatlockAgent(AgentInterface):

        # Create a fresh agent instance with scoped tools only
        # This ensures Tatlock can ONLY use tools recommended by the Steward
-        clean_host = self.ollama_host.rstrip('/')
-        base_url = f"{clean_host}/v1"
-
-        ollama_model = OpenAIChatModel(
-            model_name=self.model_name,
-            provider=get_ollama_provider()
-        )
+        model = get_model()

        # Create agent with scoped tools
        # Tools from household registry are already PydanticAI Tool objects
        scoped_agent = Agent(
-            ollama_model,
+            model,
            system_prompt=TATLOCK_SYSTEM_PROMPT,
            tools=scoped_tools,  # Pass tools directly to Agent constructor
        )
@@ -541,8 +536,7 @@ class TatlockAgent(AgentInterface):
        Yields:
            Text chunks from the streaming response
        """
-        from pydantic_ai.models.openai import OpenAIChatModel
-        from src.ollama.provider import get_ollama_provider
+        from src.anthropic.model_selector import get_model

        logger.info(
            "tatlock_run_with_scoped_tools_stream",
@@ -552,17 +546,11 @@ class TatlockAgent(AgentInterface):
        )

        # Create a fresh agent instance with scoped tools only
-        clean_host = self.ollama_host.rstrip('/')
-        base_url = f"{clean_host}/v1"
-
-        ollama_model = OpenAIChatModel(
-            model_name=self.model_name,
-            provider=get_ollama_provider()
-        )
+        model = get_model()

        # Create agent with scoped tools
        scoped_agent = Agent(
-            ollama_model,
+            model,
            system_prompt=TATLOCK_SYSTEM_PROMPT,
            tools=scoped_tools,
        )
@@ -627,7 +615,7 @@ class TatlockAgent(AgentInterface):
            steward_note: Note from Steward (invisible to user)
            scoped_tools: List of tool definitions from household registry
            message_history: Conversation history
-            tool_tracker: Optional tool call tracker for benchmarking
+            tool_tracker: Optional tool call tracker for analysis

        Returns:
            dict with:
@@ -636,8 +624,6 @@ class TatlockAgent(AgentInterface):
                - tool_outputs: Dict mapping tool names to their outputs
                - raw_output: The agent's raw text output
        """
-        from pydantic_ai.models.openai import OpenAIChatModel
-        from src.ollama.provider import get_ollama_provider
        from pydantic_ai.settings import ModelSettings
        from pydantic_ai.messages import (
            ModelRequest,
@@ -647,6 +633,7 @@ class TatlockAgent(AgentInterface):
            ToolCallPart,
            ToolReturnPart,
        )
+        from src.anthropic.model_selector import get_model

        logger.info(
            "tatlock_orchestrate_tool_calls",
@@ -655,18 +642,22 @@ class TatlockAgent(AgentInterface):
            history_length=len(message_history),
        )

-        # Create a fresh agent instance with scoped tools only
-        clean_host = self.ollama_host.rstrip('/')
-        base_url = f"{clean_host}/v1"
-
-        ollama_model = OpenAIChatModel(
-            model_name=self.model_name,
-            provider=get_ollama_provider()
+        # Start tracing span for orchestration phase
+        orchestrate_span = start_span(
+            "tatlock_orchestrate",
+            SpanType.TATLOCK,
+            metadata={
+                "scoped_tool_count": len(scoped_tools),
+                "tool_names": [getattr(t, '__name__', str(t)) for t in scoped_tools[:5]],
+            },
        )

+        # Create a fresh agent instance with scoped tools only
+        model = get_model()
+
        # Create agent with scoped tools
        scoped_agent = Agent(
-            ollama_model,
+            model,
            system_prompt=TATLOCK_SYSTEM_PROMPT,
            tools=scoped_tools,
        )
@@ -731,6 +722,23 @@ class TatlockAgent(AgentInterface):
            tool_output_count=len(tool_outputs),
        )

+        # Add tool-level spans from result messages
+        if orchestrate_span:
+            add_tool_spans_from_messages(result.new_messages(), orchestrate_span)
+
+        # End orchestration span with results
+        end_span(
+            orchestrate_span,
+            metadata_update={
+                "tools_called": tools_called,
+                "expert_count": len(expert_results),
+                "tool_output_count": len(tool_outputs),
+            },
+            details_update={
+                "steward_note_preview": steward_note[:500] if steward_note else None,
+            },
+        )
+
        return {
            "tools_called": tools_called,
            "expert_results": expert_results,
@@ -758,9 +766,8 @@ class TatlockAgent(AgentInterface):
        Returns:
            str: Butler-toned response synthesized from all results
        """
-        from pydantic_ai.models.openai import OpenAIChatModel
-        from src.ollama.provider import get_ollama_provider
        from pydantic_ai.messages import ModelRequest, ModelResponse, UserPromptPart, TextPart
+        from src.anthropic.model_selector import get_model

        logger.info(
            "tatlock_synthesize_from_results",
@@ -769,6 +776,16 @@ class TatlockAgent(AgentInterface):
            tool_count=len(orchestration_results.get("tool_outputs", {})),
        )

+        # Start tracing span for synthesis phase
+        synthesize_span = start_span(
+            "tatlock_synthesize",
+            SpanType.TATLOCK,
+            metadata={
+                "expert_count": len(orchestration_results.get("expert_results", {})),
+                "tool_output_count": len(orchestration_results.get("tool_outputs", {})),
+            },
+        )
+
        # Build synthesis prompt with all available information
        synthesis_parts = []
        synthesis_parts.append(f"The user asked: {user_message}")
@@ -789,25 +806,19 @@ class TatlockAgent(AgentInterface):
            synthesis_parts.append("")

        synthesis_parts.append(
-            "Based on this information, provide a response to the user. "
-            "Maintain your butler personality - address them as 'sir', "
-            "use formal but personable language, and be helpful."
+            "Synthesize a response for the user. Be direct and confident. "
+            "Lead with the answer - no apologies, no caveats, no 'mix-ups'. "
+            "Address them as 'sir', be concise, add dry wit if appropriate."
        )

        synthesis_prompt = "\n".join(synthesis_parts)

        # Create synthesis agent (no tools needed)
-        clean_host = self.ollama_host.rstrip('/')
-        base_url = f"{clean_host}/v1"
-
-        ollama_model = OpenAIChatModel(
-            model_name=self.model_name,
-            provider=get_ollama_provider()
-        )
+        model = get_model()

        # Synthesis agent uses butler prompt but no tools
        synthesis_agent = Agent(
-            ollama_model,
+            model,
            system_prompt=TATLOCK_SYSTEM_PROMPT,
            # No tools for synthesis phase
        )
@@ -841,6 +852,18 @@ class TatlockAgent(AgentInterface):
            response_preview=result.output[:100],
        )

+        # End synthesis span with result
+        end_span(
+            synthesize_span,
+            metadata_update={
+                "response_length": len(result.output),
+            },
+            details_update={
+                "synthesis_prompt": synthesis_prompt[:1000],
+                "response_preview": result.output[:500],
+            },
+        )
+
        return result.output

    async def get_capabilities(self) -> dict:
@@ -0,0 +1,17 @@
+"""
+Anthropic/Claude integration module.
+
+Provides model selection with automatic fallback between Claude and Ollama.
+"""
+
+from src.anthropic.model_selector import (
+    check_claude_health,
+    get_model,
+    is_claude_available,
+)
+
+__all__ = [
+    "check_claude_health",
+    "get_model",
+    "is_claude_available",
+]
@@ -0,0 +1,151 @@
+"""
+Model selector for Claude/Ollama backend switching.
+
+Provides automatic model selection with Claude as preferred backend
+and Ollama as offline fallback.
+"""
+
+from typing import Union
+
+from pydantic_ai.models.anthropic import AnthropicModel
+from pydantic_ai.models.openai import OpenAIChatModel
+
+from src.core.config import config
+from src.core.logging_config import get_logger
+
+logger = get_logger(__name__)
+
+# Cached health check result (set once at startup)
+_claude_available: bool | None = None
+
+
+async def check_claude_health() -> bool:
+    """
+    Check if Claude API is reachable and working.
+
+    This should be called once at application startup.
+    The result is cached in `_claude_available`.
+
+    Returns:
+        True if Claude API is accessible, False otherwise.
+    """
+    global _claude_available
+
+    # No API key configured - Claude not available
+    if not config.ANTHROPIC_API_KEY:
+        logger.info(
+            "claude_health_check_skipped",
+            reason="no_api_key",
+        )
+        _claude_available = False
+        return False
+
+    try:
+        from anthropic import AsyncAnthropic
+
+        client = AsyncAnthropic(api_key=config.ANTHROPIC_API_KEY)
+
+        # Minimal API call to verify connectivity
+        # Using a tiny max_tokens to minimize cost
+        await client.messages.create(
+            model=config.ANTHROPIC_MODEL,
+            max_tokens=1,
+            messages=[{"role": "user", "content": "hi"}],
+        )
+
+        _claude_available = True
+        logger.info(
+            "claude_health_check_passed",
+            model=config.ANTHROPIC_MODEL,
+        )
+        return True
+
+    except Exception as e:
+        _claude_available = False
+        logger.warning(
+            "claude_health_check_failed",
+            error=str(e),
+            model=config.ANTHROPIC_MODEL,
+        )
+        return False
+
+
+def is_claude_available() -> bool:
+    """
+    Check if Claude is available (from cached health check result).
+
+    Returns:
+        True if Claude API was reachable at startup, False otherwise.
+
+    Note:
+        Returns False if health check hasn't been run yet.
+        Call `check_claude_health()` at startup first.
+    """
+    return _claude_available is True
+
+
+def get_model(prefer_cloud: bool | None = None) -> Union[AnthropicModel, OpenAIChatModel]:
+    """
+    Get the best available model.
+
+    Returns Claude if available and preferred, otherwise Ollama.
+
+    Args:
+        prefer_cloud: Override config.PREFER_CLOUD_BACKEND for this call.
+                     If None, uses the config value.
+
+    Returns:
+        PydanticAI model instance (AnthropicModel or OpenAIChatModel).
+
+    Example:
+        >>> model = get_model()
+        >>> agent = Agent(model, system_prompt="...")
+    """
+    # Determine preference
+    use_cloud = prefer_cloud if prefer_cloud is not None else config.PREFER_CLOUD_BACKEND
+
+    # Use Claude if available and preferred
+    if use_cloud and is_claude_available():
+        logger.debug(
+            "model_selected",
+            backend="claude",
+            model=config.ANTHROPIC_MODEL,
+        )
+        return AnthropicModel(
+            model_name=config.ANTHROPIC_MODEL,
+            api_key=config.ANTHROPIC_API_KEY,
+        )
+
+    # Fall back to Ollama
+    from src.ollama.provider import get_ollama_provider
+
+    logger.debug(
+        "model_selected",
+        backend="ollama",
+        model=config.OLLAMA_DEFAULT_MODEL,
+        reason="fallback" if use_cloud else "preferred_local",
+    )
+    return OpenAIChatModel(
+        model_name=config.OLLAMA_DEFAULT_MODEL,
+        provider=get_ollama_provider(),
+    )
+
+
+def get_model_info() -> dict:
+    """
+    Get information about the current model configuration.
+
+    Useful for health checks and debugging.
+
+    Returns:
+        Dict with backend, model name, and availability info.
+    """
+    use_cloud = config.PREFER_CLOUD_BACKEND and is_claude_available()
+
+    return {
+        "backend": "claude" if use_cloud else "ollama",
+        "model": config.ANTHROPIC_MODEL if use_cloud else config.OLLAMA_DEFAULT_MODEL,
+        "claude_available": is_claude_available(),
+        "claude_configured": bool(config.ANTHROPIC_API_KEY),
+        "prefer_cloud": config.PREFER_CLOUD_BACKEND,
+    }
@@ -1,345 +0,0 @@
-"""
-Performance benchmark storage using Redis.
-
-Tracks operation timing, tool usage, and recommendation accuracy across sessions.
-Provides time-series data for performance analysis and optimization.
-"""
-import json
-from datetime import datetime, timezone
-from typing import Any, Literal, Optional
-
-import redis.asyncio as redis
-from pydantic import BaseModel, Field
-
-from .config import config
-from .logging_config import get_logger
-
-logger = get_logger(__name__)
-
-
-class PerformanceBenchmark(BaseModel):
-    """
-    Performance benchmark record.
-
-    Stores timing and metadata for operations like Steward analysis,
-    tool calls, and agent execution.
-    """
-    timestamp: datetime = Field(default_factory=lambda: datetime.now(timezone.utc))
-    operation: str  # "steward_analysis", "tool_call", "tatlock_execution"
-    duration_seconds: float
-    success: bool
-
-    # Steward-specific fields
-    recommendation_count: Optional[int] = None
-    confidence: Optional[float] = None
-
-    # Tool-specific fields
-    tool_name: Optional[str] = None
-    was_recommended: Optional[bool] = None
-    was_actually_used: Optional[bool] = None
-
-    # Context
-    conversation_id: Optional[str] = None
-    metadata: dict[str, Any] = Field(default_factory=dict)
-
-    def to_redis_dict(self) -> dict[str, Any]:
-        """Convert to dict suitable for Redis storage."""
-        data = self.model_dump()
-        data["timestamp"] = self.timestamp.isoformat()
-        data["metadata"] = json.dumps(self.metadata)
-        # Convert booleans to strings (Redis doesn't accept bool type)
-        for key, value in data.items():
-            if isinstance(value, bool):
-                data[key] = str(value)
-        return data
-
-    @classmethod
-    def from_redis_dict(cls, data: dict[str, Any]) -> "PerformanceBenchmark":
-        """Reconstruct from Redis dict."""
-        data["timestamp"] = datetime.fromisoformat(data["timestamp"])
-        data["metadata"] = json.loads(data.get("metadata", "{}"))
-        # Convert string booleans back to bool
-        for key in ["success", "was_recommended", "was_actually_used"]:
-            if key in data and isinstance(data[key], str):
-                data[key] = data[key] == "True"
-        return cls(**data)
-
-
-class BenchmarkStore:
-    """
-    Redis-backed benchmark storage with automatic expiry.
-
-    Stores performance metrics in time-series format with 30-day retention.
-    Provides querying capabilities for analysis and reporting.
-    """
-
-    def __init__(self, redis_client: Optional[redis.Redis] = None):
-        """
-        Initialize benchmark store.
-
-        Args:
-            redis_client: Optional Redis client. If None, creates from config.
-        """
-        self._client = redis_client
-        self._ttl_days = 30  # 30-day retention
-
-    async def _get_client(self) -> redis.Redis:
-        """Get or create Redis client."""
-        if self._client is None:
-            self._client = redis.from_url(
-                config.redis_url,
-                encoding="utf-8",
-                decode_responses=True,
-                socket_timeout=config.REDIS_TIMEOUT,
-                socket_connect_timeout=config.REDIS_TIMEOUT,
-            )
-        return self._client
-
-    async def record(self, benchmark: PerformanceBenchmark) -> None:
-        """
-        Record a performance benchmark.
-
-        Args:
-            benchmark: Performance benchmark to record
-
-        Example:
-            >>> await store.record(PerformanceBenchmark(
-            ...     operation="steward_analysis",
-            ...     duration_seconds=1.23,
-            ...     success=True,
-            ...     recommendation_count=3,
-            ... ))
-        """
-        if not config.ENABLE_BENCHMARKS:
-            return
-
-        try:
-            client = await self._get_client()
-
-            # Generate key: benchmark:{operation}:{timestamp_ms}
-            timestamp_ms = int(benchmark.timestamp.timestamp() * 1000)
-            key = f"benchmark:{benchmark.operation}:{timestamp_ms}"
-
-            # Store as hash
-            await client.hset(key, mapping=benchmark.to_redis_dict())
-
-            # Set expiry
-            await client.expire(key, self._ttl_days * 24 * 60 * 60)
-
-            # Add to sorted set for time-based queries
-            index_key = f"benchmark_index:{benchmark.operation}"
-            await client.zadd(index_key, {key: timestamp_ms})
-            await client.expire(index_key, self._ttl_days * 24 * 60 * 60)
-
-            logger.debug(
-                "benchmark_recorded",
-                operation=benchmark.operation,
-                duration=benchmark.duration_seconds,
-                success=benchmark.success,
-            )
-
-        except Exception as e:
-            logger.warning(
-                "benchmark_recording_failed",
-                error=str(e),
-                operation=benchmark.operation,
-            )
-            # Don't fail the request if benchmarking fails
-
-    async def query(
-        self,
-        operation: str,
-        start_time: Optional[datetime] = None,
-        end_time: Optional[datetime] = None,
-        limit: int = 100,
-    ) -> list[PerformanceBenchmark]:
-        """
-        Query benchmarks by operation and time range.
-
-        Args:
-            operation: Operation name to filter by
-            start_time: Start of time range (inclusive)
-            end_time: End of time range (inclusive)
-            limit: Maximum number of results
-
-        Returns:
-            List of benchmarks matching the query
-
-        Example:
-            >>> from datetime import timedelta
-            >>> now = datetime.now(timezone.utc)
-            >>> yesterday = now - timedelta(days=1)
-            >>> benchmarks = await store.query(
-            ...     "steward_analysis",
-            ...     start_time=yesterday,
-            ...     limit=50
-            ... )
-        """
-        if not config.ENABLE_BENCHMARKS:
-            return []
-
-        try:
-            client = await self._get_client()
-            index_key = f"benchmark_index:{operation}"
-
-            # Convert time range to timestamps
-            min_score = (
-                int(start_time.timestamp() * 1000)
-                if start_time
-                else "-inf"
-            )
-            max_score = (
-                int(end_time.timestamp() * 1000)
-                if end_time
-                else "+inf"
-            )
-
-            # Query sorted set
-            keys = await client.zrevrangebyscore(
-                index_key,
-                max_score,
-                min_score,
-                start=0,
-                num=limit,
-            )
-
-            # Fetch benchmark data
-            benchmarks = []
-            for key in keys:
-                data = await client.hgetall(key)
-                if data:
-                    benchmarks.append(PerformanceBenchmark.from_redis_dict(data))
-
-            return benchmarks
-
-        except Exception as e:
-            logger.error(
-                "benchmark_query_failed",
-                error=str(e),
-                operation=operation,
-            )
-            return []
-
-    async def get_statistics(
-        self,
-        operation: str,
-        start_time: Optional[datetime] = None,
-        end_time: Optional[datetime] = None,
-    ) -> dict[str, Any]:
-        """
-        Get aggregate statistics for an operation.
-
-        Args:
-            operation: Operation name
-            start_time: Start of time range
-            end_time: End of time range
-
-        Returns:
-            Dictionary with statistics (count, avg_duration, success_rate, etc.)
-
-        Example:
-            >>> stats = await store.get_statistics("steward_analysis")
-            >>> print(f"Average duration: {stats['avg_duration']}s")
-            >>> print(f"Success rate: {stats['success_rate']}%")
-        """
-        benchmarks = await self.query(operation, start_time, end_time, limit=1000)
-
-        if not benchmarks:
-            return {
-                "count": 0,
-                "avg_duration": 0.0,
-                "min_duration": 0.0,
-                "max_duration": 0.0,
-                "success_rate": 0.0,
-            }
-
-        durations = [b.duration_seconds for b in benchmarks]
-        successes = sum(1 for b in benchmarks if b.success)
-
-        return {
-            "count": len(benchmarks),
-            "avg_duration": sum(durations) / len(durations),
-            "min_duration": min(durations),
-            "max_duration": max(durations),
-            "success_rate": (successes / len(benchmarks)) * 100,
-            "total_successes": successes,
-            "total_failures": len(benchmarks) - successes,
-        }
-
-    async def get_tool_accuracy(
-        self,
-        start_time: Optional[datetime] = None,
-        end_time: Optional[datetime] = None,
-    ) -> dict[str, Any]:
-        """
-        Analyze tool recommendation accuracy.
-
-        Compares recommended tools vs actually used tools to measure
-        Steward's recommendation precision.
-
-        Args:
-            start_time: Start of time range
-            end_time: End of time range
-
-        Returns:
-            Dictionary with accuracy metrics
-
-        Example:
-            >>> accuracy = await store.get_tool_accuracy()
-            >>> print(f"Precision: {accuracy['precision']}%")
-        """
-        tool_calls = await self.query("tool_call", start_time, end_time, limit=1000)
-
-        if not tool_calls:
-            return {
-                "total_calls": 0,
-                "recommended_and_used": 0,
-                "recommended_not_used": 0,
-                "not_recommended_but_used": 0,
-                "precision": 0.0,
-            }
-
-        recommended_and_used = sum(
-            1 for b in tool_calls
-            if b.was_recommended and b.was_actually_used
-        )
-        not_recommended_but_used = sum(
-            1 for b in tool_calls
-            if not b.was_recommended and b.was_actually_used
-        )
-
-        total_used = sum(1 for b in tool_calls if b.was_actually_used)
-        precision = (
-            (recommended_and_used / total_used * 100) if total_used > 0 else 0.0
-        )
-
-        return {
-            "total_calls": len(tool_calls),
-            "total_used": total_used,
-            "recommended_and_used": recommended_and_used,
-            "not_recommended_but_used": not_recommended_but_used,
-            "precision": precision,
-        }
-
-    async def close(self) -> None:
-        """Close Redis connection."""
-        if self._client:
-            await self._client.aclose()
-            self._client = None
-
-
-# Global benchmark store instance
-_benchmark_store: Optional[BenchmarkStore] = None
-
-
-def get_benchmark_store() -> BenchmarkStore:
-    """
-    Get global benchmark store instance.
-
-    Returns:
-        BenchmarkStore instance
-    """
-    global _benchmark_store
-    if _benchmark_store is None:
-        _benchmark_store = BenchmarkStore()
-    return _benchmark_store
@@ -64,7 +64,21 @@ class Config(BaseSettings):
    API_PORT: int = Field(default=8000, description="API port")
    API_PREFIX: str = Field(default="/v1", description="API route prefix")
    
-    # Ollama Configuration
+    # Anthropic Configuration (Claude - preferred backend)
+    ANTHROPIC_API_KEY: str | None = Field(
+        default=None,
+        description="Anthropic API key for Claude access"
+    )
+    ANTHROPIC_MODEL: str = Field(
+        default="claude-sonnet-4-20250514",
+        description="Claude model to use"
+    )
+    PREFER_CLOUD_BACKEND: bool = Field(
+        default=True,
+        description="Prefer Claude over Ollama when available"
+    )
+
+    # Ollama Configuration (local fallback)
    OLLAMA_HOST: HttpUrl = Field(
        default="http://localhost:11434",
        description="Ollama server URL"
@@ -101,10 +115,6 @@ class Config(BaseSettings):
        default=6379,
        description="Redis server port"
    )
-    REDIS_BENCHMARK_DB: int = Field(
-        default=6,
-        description="Redis database number for benchmarks"
-    )
    REDIS_TIMEOUT: int = Field(
        default=5,
        description="Redis connection timeout in seconds"
@@ -158,7 +168,7 @@ class Config(BaseSettings):
        description="Ollama model for embeddings"
    )

-    # Redis Memory Database (separate from benchmarks)
+    # Redis Memory Database
    REDIS_MEMORY_DB: int = Field(
        default=1,
        description="Redis database number for memory cache"
@@ -173,7 +183,6 @@ class Config(BaseSettings):
        default=None,
        description="Logging level (auto-set based on environment if not specified)"
    )
-    ENABLE_BENCHMARKS: bool = Field(default=True, description="Enable performance benchmarking")

    # User Configuration
    DEFAULT_USER: str | None = Field(
@@ -190,11 +199,6 @@ class Config(BaseSettings):
    CORS_ALLOW_METHODS: list[str] = ["*"]
    CORS_ALLOW_HEADERS: list[str] = ["*"]

-    @property
-    def redis_url(self) -> str:
-        """Construct Redis connection URL for benchmarks."""
-        return f"redis://{self.REDIS_HOST}:{self.REDIS_PORT}/{self.REDIS_BENCHMARK_DB}"
-
    @property
    def redis_memory_url(self) -> str:
        """Construct Redis connection URL for memory cache."""
@@ -6,7 +6,7 @@ Provides short-term memory storage with TTL:
 - Recent entities mentioned in conversation
 - User-scoped with conversation isolation

-Uses Redis DB 2 (separate from benchmarks in DB 1).
+Uses Redis DB 1.
 """
 import json
 from typing import Any
@@ -11,6 +11,7 @@ from src.agents.steward import analyze_request, format_steward_note
 from src.agents.steward.schemas import StewardRecommendation
 from src.core.household_registry import get_household_registry
 from src.core.logging_config import get_logger
+from src.core.tracing import trace_span, SpanType

 logger = get_logger(__name__)

@@ -93,12 +94,32 @@ async def preprocess_request(
        conversation_id=conversation_id,
    )

-    # Call Steward with full conversation history
-    recommendation = await analyze_request(
-        enriched_request,
-        conversation_history=conversation_history,
-        conversation_id=conversation_id,
-    )
+    # Call Steward with full conversation history (traced)
+    async with trace_span(
+        "steward_analysis",
+        SpanType.STEWARD,
+        metadata={
+            "request_preview": user_request[:100],
+            "history_length": len(conversation_history),
+        },
+    ) as span:
+        recommendation = await analyze_request(
+            enriched_request,
+            conversation_history=conversation_history,
+            conversation_id=conversation_id,
+        )
+
+        # Update span with results
+        if span:
+            span.metadata.update({
+                "recommended_capabilities": recommendation.recommended_capabilities,
+                "complexity": recommendation.estimated_complexity,
+                "has_memory_context": bool(recommendation.memory_context),
+                "has_conversation_context": recommendation.conversation_context.has_previous_context,
+            })
+            span.details["reasoning"] = recommendation.reasoning
+            if recommendation.enriched_query:
+                span.details["enriched_query"] = recommendation.enriched_query

    # Format note for Tatlock (includes conversation context)
    steward_note = await format_steward_note(recommendation)
@@ -9,6 +9,7 @@ from src.agents.biographer import register_biographer
 from src.agents.housekeeper import register_housekeeper
 from src.agents.librarian import register_librarian
 from src.agents.tatlock_core import TATLOCK_CORE_CAPABILITY, tatlock_core_tools
+from src.anthropic.model_selector import check_claude_health, get_model_info
 from src.core.household_registry import get_household_registry
 from src.core.logging_config import get_logger

@@ -81,19 +82,29 @@ def register_household_members():
    )


-def initialize_application():
+async def initialize_application():
    """
    Initialize the application.

    Performs all startup tasks:
-    1. Register household members
-    2. (Future) Initialize connections
-    3. (Future) Load configuration
+    1. Check Claude API health (for backend selection)
+    2. Register household members
+    3. (Future) Initialize connections

    This should be called once during application startup.
    """
    logger.info("application_initialization_starting")

+    # Check Claude API health for backend selection
+    await check_claude_health()
+    model_info = get_model_info()
+    logger.info(
+        "model_backend_configured",
+        backend=model_info["backend"],
+        model=model_info["model"],
+        claude_available=model_info["claude_available"],
+    )
+
    # Register household members
    register_household_members()

@@ -1,13 +1,11 @@
 """
-Tool call tracking and benchmarking.
+Tool call tracking.

 Tracks which tools are recommended by the Steward versus which tools
-are actually used by Tatlock, recording benchmarks for analysis.
+are actually used by Tatlock for debugging and analysis.
 """
-from datetime import datetime, timezone
 from typing import Optional

-from src.core.benchmarks import PerformanceBenchmark, get_benchmark_store
 from src.core.logging_config import get_logger

 logger = get_logger(__name__)
@@ -15,7 +13,7 @@ logger = get_logger(__name__)

 class ToolCallTracker:
    """
-    Tracks tool calls for benchmarking and accuracy analysis.
+    Tracks tool calls for accuracy analysis.

    Compares Steward's recommendations with Tatlock's actual tool usage
    to measure recommendation accuracy.
@@ -53,6 +51,10 @@ class ToolCallTracker:
            return tool_name.replace("delegate_to_", "")
        return tool_name

+    def log_call(self, message: str):
+        """Log a tool call message (for UI display)."""
+        logger.debug("tool_call_message", message=message)
+
    async def track_call(self, tool_name: str, duration: float):
        """
        Record a tool call with timing.
@@ -78,23 +80,6 @@ class ToolCallTracker:
                recommended=list(self.recommended_capabilities),
            )

-        # Record benchmark to Redis
-        benchmark = PerformanceBenchmark(
-            timestamp=datetime.now(timezone.utc),
-            operation="tool_call",
-            duration_seconds=duration,
-            success=True,  # If we got here, the call succeeded
-            tool_name=tool_name,
-            was_recommended=was_recommended,
-            was_actually_used=True,
-            conversation_id=self.conversation_id,
-            metadata={
-                "recommended_capabilities": list(self.recommended_capabilities),
-            },
-        )
-
-        await get_benchmark_store().record(benchmark)
-
        logger.debug(
            "tool_call_tracked",
            tool_name=tool_name,
@@ -124,24 +109,6 @@ class ToolCallTracker:
                conversation_id=self.conversation_id,
            )

-            # Record benchmarks for unused recommendations
-            for tool_name in unused_tools:
-                benchmark = PerformanceBenchmark(
-                    timestamp=datetime.now(timezone.utc),
-                    operation="tool_call",
-                    duration_seconds=0.0,  # Not used
-                    success=True,
-                    tool_name=tool_name,
-                    was_recommended=True,
-                    was_actually_used=False,
-                    conversation_id=self.conversation_id,
-                    metadata={
-                        "recommended_capabilities": list(self.recommended_capabilities),
-                        "reason": "recommended_but_unused",
-                    },
-                )
-                await get_benchmark_store().record(benchmark)
-
        # Log summary
        total_calls = sum(len(durations) for durations in self.actual_calls.values())
        logger.info(
@@ -0,0 +1,434 @@
+"""
+Lightweight request tracing for local development.
+
+Captures the full request flow through Tatlock's multi-agent architecture
+as structured JSON traces for debugging and optimization.
+
+Enable via DEBUG=true environment variable.
+
+Traces are written to logs/traces/{trace_id}.json
+View with logs/traces/viewer.html
+"""
+from contextlib import asynccontextmanager
+from contextvars import ContextVar
+from dataclasses import dataclass, field
+from datetime import datetime, timezone
+from enum import Enum
+from pathlib import Path
+from typing import Any
+
+import json
+import secrets
+
+from src.core.logging_config import get_logger
+
+logger = get_logger(__name__)
+
+
+class SpanType(str, Enum):
+    """Types of traced operations."""
+    ROUTER = "router"
+    STEWARD = "steward"
+    TATLOCK = "tatlock"
+    EXPERT = "expert"
+    TOOL = "tool"
+
+
+class SpanStatus(str, Enum):
+    """Span completion status."""
+    OK = "ok"
+    ERROR = "error"
+
+
+@dataclass
+class Span:
+    """A single traced operation."""
+    span_id: str
+    name: str
+    type: SpanType
+    start_time: datetime
+    parent_id: str | None = None
+    end_time: datetime | None = None
+    status: SpanStatus = SpanStatus.OK
+    metadata: dict[str, Any] = field(default_factory=dict)
+    details: dict[str, Any] = field(default_factory=dict)
+    children: list[str] = field(default_factory=list)
+    error: str | None = None
+
+    @property
+    def duration_ms(self) -> float | None:
+        """Calculate duration in milliseconds."""
+        if self.end_time and self.start_time:
+            return (self.end_time - self.start_time).total_seconds() * 1000
+        return None
+
+    def to_dict(self) -> dict[str, Any]:
+        """Convert span to dictionary for JSON serialization."""
+        result = {
+            "span_id": self.span_id,
+            "parent_id": self.parent_id,
+            "name": self.name,
+            "type": self.type.value,
+            "start_time": self.start_time.isoformat(),
+            "end_time": self.end_time.isoformat() if self.end_time else None,
+            "duration_ms": round(self.duration_ms, 2) if self.duration_ms else None,
+            "status": self.status.value,
+            "metadata": self.metadata if self.metadata else None,
+        }
+        # Only include non-empty optional fields
+        if self.details:
+            result["details"] = self.details
+        if self.children:
+            result["children"] = self.children
+        if self.error:
+            result["error"] = self.error
+        return {k: v for k, v in result.items() if v is not None}
+
+
+@dataclass
+class Trace:
+    """Complete trace of a request."""
+    trace_id: str
+    conversation_id: str | None
+    user: str
+    timestamp: datetime
+    request: dict[str, Any]
+    spans: list[Span] = field(default_factory=list)
+    response: dict[str, Any] | None = None
+    status: str = "in_progress"
+
+    @property
+    def total_duration_ms(self) -> float | None:
+        """Calculate total trace duration from span timings."""
+        if not self.spans:
+            return None
+        start = min(s.start_time for s in self.spans)
+        ends = [s.end_time for s in self.spans if s.end_time]
+        if not ends:
+            return None
+        end = max(ends)
+        return (end - start).total_seconds() * 1000
+
+    def to_dict(self) -> dict[str, Any]:
+        """Convert trace to dictionary for JSON serialization."""
+        return {
+            "trace_id": self.trace_id,
+            "conversation_id": self.conversation_id,
+            "user": self.user,
+            "timestamp": self.timestamp.isoformat(),
+            "total_duration_ms": round(self.total_duration_ms, 2) if self.total_duration_ms else None,
+            "status": self.status,
+            "request": self.request,
+            "response": self.response,
+            "spans": [s.to_dict() for s in self.spans],
+        }
+
+
+# ContextVar for async-safe trace propagation
+_current_trace: ContextVar[Trace | None] = ContextVar("current_trace", default=None)
+_current_span: ContextVar[Span | None] = ContextVar("current_span", default=None)
+
+
+def tracing_enabled() -> bool:
+    """Check if tracing is enabled (requires DEBUG=true)."""
+    from src.core.config import config
+    return config.DEBUG
+
+
+def _generate_id(prefix: str = "") -> str:
+    """Generate unique ID with optional prefix."""
+    return f"{prefix}{secrets.token_hex(8)}"
+
+
+def start_trace(
+    conversation_id: str | None,
+    user: str,
+    request: dict[str, Any],
+) -> Trace | None:
+    """
+    Start a new trace for a request.
+
+    Args:
+        conversation_id: Conversation identifier
+        user: User identifier
+        request: Request data (should include preview and full)
+
+    Returns:
+        Trace object if tracing enabled, None otherwise
+    """
+    if not tracing_enabled():
+        return None
+
+    trace = Trace(
+        trace_id=_generate_id("trace_"),
+        conversation_id=conversation_id,
+        user=user,
+        timestamp=datetime.now(timezone.utc),
+        request=request,
+    )
+    _current_trace.set(trace)
+
+    logger.debug("trace_started", trace_id=trace.trace_id, user=user)
+    return trace
+
+
+def get_current_trace() -> Trace | None:
+    """Get the current trace from context."""
+    return _current_trace.get()
+
+
+def get_current_span() -> Span | None:
+    """Get the current span from context."""
+    return _current_span.get()
+
+
+def start_span(
+    name: str,
+    span_type: SpanType,
+    metadata: dict[str, Any] | None = None,
+    details: dict[str, Any] | None = None,
+) -> Span | None:
+    """
+    Start a new span within the current trace.
+
+    Args:
+        name: Span name (e.g., "steward_analysis")
+        span_type: Type of operation
+        metadata: Quick-access metadata (shown in timeline)
+        details: Expandable details (prompts, full responses)
+
+    Returns:
+        Span object if tracing enabled, None otherwise
+    """
+    trace = get_current_trace()
+    if not trace:
+        return None
+
+    parent = get_current_span()
+    span = Span(
+        span_id=_generate_id("span_"),
+        name=name,
+        type=span_type,
+        start_time=datetime.now(timezone.utc),
+        parent_id=parent.span_id if parent else None,
+        metadata=metadata or {},
+        details=details or {},
+    )
+
+    # Add to parent's children list
+    if parent:
+        parent.children.append(span.span_id)
+
+    trace.spans.append(span)
+    _current_span.set(span)
+
+    logger.debug(
+        "span_started",
+        span_id=span.span_id,
+        name=name,
+        type=span_type.value,
+        parent_id=span.parent_id,
+    )
+    return span
+
+
+def end_span(
+    span: Span | None = None,
+    status: SpanStatus = SpanStatus.OK,
+    metadata_update: dict[str, Any] | None = None,
+    details_update: dict[str, Any] | None = None,
+    error: str | None = None,
+) -> None:
+    """
+    End a span and restore parent as current.
+
+    Args:
+        span: Span to end (defaults to current span)
+        status: Completion status
+        metadata_update: Additional metadata to merge
+        details_update: Additional details to merge
+        error: Error message if failed
+    """
+    if span is None:
+        span = get_current_span()
+    if not span:
+        return
+
+    span.end_time = datetime.now(timezone.utc)
+    span.status = status
+    if error:
+        span.error = error
+        span.status = SpanStatus.ERROR
+    if metadata_update:
+        span.metadata.update(metadata_update)
+    if details_update:
+        span.details.update(details_update)
+
+    # Restore parent span as current
+    trace = get_current_trace()
+    if trace and span.parent_id:
+        parent = next((s for s in trace.spans if s.span_id == span.parent_id), None)
+        _current_span.set(parent)
+    else:
+        _current_span.set(None)
+
+    logger.debug(
+        "span_ended",
+        span_id=span.span_id,
+        duration_ms=span.duration_ms,
+        status=status.value,
+    )
+
+
+def end_trace(
+    response: dict[str, Any] | None = None,
+    status: str = "completed",
+) -> str | None:
+    """
+    End the current trace and write to file.
+
+    Args:
+        response: Response data to include
+        status: Final trace status ("completed" or "error")
+
+    Returns:
+        Path to trace file if written, None otherwise
+    """
+    trace = get_current_trace()
+    if not trace:
+        return None
+
+    trace.response = response
+    trace.status = status
+
+    # Write trace to file
+    trace_path = _write_trace(trace)
+
+    # Clear context
+    _current_trace.set(None)
+    _current_span.set(None)
+
+    logger.info(
+        "trace_completed",
+        trace_id=trace.trace_id,
+        total_duration_ms=round(trace.total_duration_ms, 2) if trace.total_duration_ms else None,
+        span_count=len(trace.spans),
+        path=str(trace_path) if trace_path else None,
+    )
+
+    return str(trace_path) if trace_path else None
+
+
+def _write_trace(trace: Trace) -> Path | None:
+    """Write trace to JSON file."""
+    try:
+        # Ensure traces directory exists
+        traces_dir = Path("logs/traces")
+        traces_dir.mkdir(parents=True, exist_ok=True)
+
+        # Write trace file
+        trace_path = traces_dir / f"{trace.trace_id}.json"
+        with open(trace_path, "w") as f:
+            json.dump(trace.to_dict(), f, indent=2, default=str)
+
+        return trace_path
+
+    except Exception as e:
+        logger.error("trace_write_failed", error=str(e), trace_id=trace.trace_id)
+        return None
+
+
+@asynccontextmanager
+async def trace_span(
+    name: str,
+    span_type: SpanType,
+    metadata: dict[str, Any] | None = None,
+    details: dict[str, Any] | None = None,
+):
+    """
+    Async context manager for tracing a span.
+
+    Automatically handles start/end timing and error capture.
+
+    Usage:
+        async with trace_span("steward_analysis", SpanType.STEWARD) as span:
+            result = await analyze_request(...)
+            if span:
+                span.metadata["result_count"] = len(result)
+
+    Args:
+        name: Span name
+        span_type: Type of operation
+        metadata: Initial metadata
+        details: Initial details (expandable in viewer)
+
+    Yields:
+        Span object or None if tracing disabled
+    """
+    span = start_span(name, span_type, metadata, details)
+    try:
+        yield span
+    except Exception as e:
+        end_span(span, SpanStatus.ERROR, error=str(e))
+        raise
+    else:
+        end_span(span, SpanStatus.OK)
+
+
+def add_tool_spans_from_messages(messages: list[Any], parent_span: Span | None = None) -> None:
+    """
+    Extract tool calls from PydanticAI result messages and add as child spans.
+
+    Call this after an agent.run() to capture tool-level timing retroactively.
+    Note: Since we don't have actual timing, we estimate based on sequence.
+
+    Args:
+        messages: List from result.new_messages()
+        parent_span: Parent span to attach tool spans to
+    """
+    trace = get_current_trace()
+    if not trace or not parent_span:
+        return
+
+    # Import PydanticAI message types
+    try:
+        from pydantic_ai.messages import ModelRequest, ModelResponse, ToolCallPart, ToolReturnPart
+    except ImportError:
+        return
+
+    # Track tool calls and their returns
+    tool_calls: dict[str, dict[str, Any]] = {}
+
+    for msg in messages:
+        if isinstance(msg, ModelResponse):
+            for part in msg.parts:
+                if isinstance(part, ToolCallPart):
+                    tool_calls[part.tool_call_id] = {
+                        "name": part.tool_name,
+                        "args": part.args if hasattr(part, 'args') else {},
+                    }
+        elif isinstance(msg, ModelRequest):
+            for part in msg.parts:
+                if isinstance(part, ToolReturnPart):
+                    if part.tool_call_id in tool_calls:
+                        tool_info = tool_calls[part.tool_call_id]
+                        # Create a span for this tool call
+                        span = Span(
+                            span_id=_generate_id("span_"),
+                            name=tool_info["name"],
+                            type=SpanType.TOOL,
+                            start_time=parent_span.start_time,  # Approximate
+                            end_time=parent_span.end_time or datetime.now(timezone.utc),
+                            parent_id=parent_span.span_id,
+                            status=SpanStatus.OK,
+                            metadata={
+                                "tool_name": tool_info["name"],
+                                "args_preview": str(tool_info.get("args", {}))[:100],
+                            },
+                            details={
+                                "args": tool_info.get("args", {}),
+                                "result": part.content[:2000] if isinstance(part.content, str) else str(part.content)[:2000],
+                            },
+                        )
+                        parent_span.children.append(span.span_id)
+                        trace.spans.append(span)
@@ -0,0 +1,153 @@
+"""
+Trace viewer router.
+
+Serves the trace viewer UI and trace files when tracing is enabled.
+Only available when DEBUG=true.
+"""
+from pathlib import Path
+
+from fastapi import APIRouter, HTTPException
+from fastapi.responses import HTMLResponse, JSONResponse
+
+from src.core.config import config
+from src.core.logging_config import get_logger
+
+logger = get_logger(__name__)
+
+router = APIRouter(prefix="/traces", tags=["traces"])
+
+TRACES_DIR = Path("logs/traces")
+VIEWER_PATH = TRACES_DIR / "viewer.html"
+
+
+def tracing_enabled() -> bool:
+    """Check if tracing is enabled."""
+    return config.DEBUG
+
+
+@router.get("", response_class=HTMLResponse)
+async def get_trace_viewer():
+    """
+    Serve the trace viewer UI.
+
+    Returns the standalone HTML viewer for browsing traces.
+    """
+    if not tracing_enabled():
+        raise HTTPException(status_code=404, detail="Tracing not enabled")
+
+    if not VIEWER_PATH.exists():
+        raise HTTPException(status_code=404, detail="Viewer not found")
+
+    return HTMLResponse(content=VIEWER_PATH.read_text())
+
+
+@router.get("/list")
+async def list_traces(
+    limit: int = 50,
+    since_minutes: int | None = None,
+    status: str | None = None,
+    search: str | None = None,
+):
+    """
+    List available trace files.
+
+    Returns most recent traces first, with basic metadata.
+
+    Args:
+        limit: Maximum number of traces to return (default 50)
+        since_minutes: Only return traces from the last N minutes
+        status: Filter by status (completed, error, streaming)
+        search: Search in request preview text
+    """
+    if not tracing_enabled():
+        raise HTTPException(status_code=404, detail="Tracing not enabled")
+
+    if not TRACES_DIR.exists():
+        return {"traces": [], "total": 0}
+
+    import json
+    from datetime import datetime, timezone, timedelta
+
+    # Calculate cutoff time if filtering by time
+    cutoff_time = None
+    if since_minutes:
+        cutoff_time = datetime.now(timezone.utc) - timedelta(minutes=since_minutes)
+
+    # Get all trace files, sorted by modification time (newest first)
+    trace_files = sorted(
+        TRACES_DIR.glob("trace_*.json"),
+        key=lambda p: p.stat().st_mtime,
+        reverse=True,
+    )
+
+    traces = []
+    for path in trace_files:
+        if len(traces) >= limit:
+            break
+
+        try:
+            with open(path) as f:
+                data = json.load(f)
+
+            # Parse timestamp for filtering
+            trace_timestamp = data.get("timestamp")
+            if cutoff_time and trace_timestamp:
+                try:
+                    ts = datetime.fromisoformat(trace_timestamp.replace('Z', '+00:00'))
+                    if ts < cutoff_time:
+                        continue
+                except (ValueError, TypeError):
+                    pass
+
+            # Filter by status
+            trace_status = data.get("status", "")
+            if status and trace_status != status:
+                continue
+
+            # Filter by search text
+            request_preview = data.get("request", {}).get("input_preview", "")
+            if search and search.lower() not in request_preview.lower():
+                continue
+
+            traces.append({
+                "trace_id": data.get("trace_id"),
+                "timestamp": trace_timestamp,
+                "user": data.get("user"),
+                "status": trace_status,
+                "total_duration_ms": data.get("total_duration_ms"),
+                "span_count": len(data.get("spans", [])),
+                "request_preview": request_preview[:100],
+            })
+        except Exception as e:
+            logger.warning("trace_list_parse_error", path=str(path), error=str(e))
+
+    return {"traces": traces, "total": len(traces)}
+
+
+@router.get("/{trace_id}")
+async def get_trace(trace_id: str):
+    """
+    Get a specific trace by ID.
+
+    Returns the full trace JSON.
+    """
+    if not tracing_enabled():
+        raise HTTPException(status_code=404, detail="Tracing not enabled")
+
+    # Sanitize trace_id to prevent path traversal
+    if not trace_id.startswith("trace_") or "/" in trace_id or "\\" in trace_id:
+        raise HTTPException(status_code=400, detail="Invalid trace ID")
+
+    trace_path = TRACES_DIR / f"{trace_id}.json"
+
+    if not trace_path.exists():
+        raise HTTPException(status_code=404, detail="Trace not found")
+
+    try:
+        import json
+        with open(trace_path) as f:
+            data = json.load(f)
+        return JSONResponse(content=data)
+    except Exception as e:
+        logger.error("trace_read_error", trace_id=trace_id, error=str(e))
+        raise HTTPException(status_code=500, detail="Failed to read trace")
@@ -23,6 +23,7 @@ from src.core.exceptions import AppException
 from src.core.logging_config import get_logger
 from src.core.router import router as core_router
 from src.core.startup import initialize_application
+from src.core.tracing_router import router as tracing_router
 from src.models.router import router as models_router
 from src.responses.router import router as responses_router

@@ -43,14 +44,16 @@ async def lifespan(app: FastAPI) -> AsyncGenerator[None, None]:
        app_name=config.APP_NAME,
        version=config.APP_VERSION,
        environment=config.ENVIRONMENT.value,
+        prefer_cloud=config.PREFER_CLOUD_BACKEND,
+        anthropic_model=config.ANTHROPIC_MODEL,
        ollama_host=str(config.OLLAMA_HOST),
        ollama_model=config.OLLAMA_DEFAULT_MODEL,
-        redis_url=config.redis_url,
+        redis_url=config.redis_memory_url,
        log_format=config.log_format,
    )

-    # Initialize application (register household members, etc.)
-    initialize_application()
+    # Initialize application (check Claude health, register household members, etc.)
+    await initialize_application()

    yield

@@ -90,7 +93,12 @@ def create_application() -> FastAPI:
    application.include_router(chat_router, prefix=config.API_PREFIX)
    application.include_router(models_router, prefix=config.API_PREFIX)
    application.include_router(responses_router, prefix=config.API_PREFIX)  # Responses API
-    
+
+    # Conditionally include tracing router (only in debug mode)
+    if config.DEBUG:
+        application.include_router(tracing_router)
+        logger.info("tracing_router_enabled")
+
    return application


@@ -10,7 +10,6 @@ from sse_starlette.sse import EventSourceResponse
 from src.responses import service
 from src.responses.schemas import ResponseRequest, Response
 from src.core.exceptions import ModelNotFoundError, AppException
-from src.core.context import current_user, current_conversation, get_default_user
 from src.core.logging_config import get_logger

 logger = get_logger(__name__)
@@ -37,77 +36,16 @@ async def create_response(

    Returns:
        Response object or SSE stream
-
-    Example non-streaming request:
-        POST /v1/responses
-        {
-            "model": "lorem-tester",
-            "input": [{"role": "user", "content": "Hello"}],
-            "reasoning": {"effort": "medium", "summary": "auto"},
-            "stream": false
-        }
-
-    Example streaming request:
-        POST /v1/responses
-        {
-            "model": "lorem-tester",
-            "input": [{"role": "user", "content": "Hello"}],
-            "stream": true
-        }
-
-    Response format (non-streaming):
-        {
-            "id": "resp_...",
-            "object": "response",
-            "created_at": 1733529600,
-            "model": "lorem-tester",
-            "status": "completed",
-            "output": [
-                {
-                    "type": "reasoning",
-                    "id": "rs_...",
-                    "summary": ["Analyzing...", "Considering..."]
-                },
-                {
-                    "type": "message",
-                    "id": "msg_...",
-                    "role": "assistant",
-                    "content": [{"type": "output_text", "text": "Lorem ipsum..."}]
-                }
-            ],
-            "usage": {
-                "input_tokens": 10,
-                "output_tokens": 50,
-                "reasoning_tokens": 20,
-                "total_tokens": 80
-            }
-        }
-
-    Streaming format (SSE):
-        event: response.reasoning_summary_text.delta
-        data: {"delta": "Analyzing..."}
-
-        event: response.output_text.delta
-        data: {"delta": "Lorem"}
-
-        event: response.done
-        data: {"response": {...}}
    """
-    # Set request context (propagates through all async calls)
-    effective_user = request.user or get_default_user()
-    user_token = current_user.set(effective_user)
-    conv_id = request.metadata.get("conversation_id") if request.metadata else None
-    conv_token = current_conversation.set(conv_id)
-
    logger.info(
        "response_request_received",
        model=request.model,
-        user=effective_user,
-        conversation_id=conv_id,
+        user=request.user,
+        streaming=request.stream,
    )

    try:
-        # Check if this is a Tatlock request - use Steward preprocessing (Phase 2)
+        # Check if this is a Tatlock request - use Steward preprocessing
        model_id = request.model
        if "." in model_id:
            model_id = model_id.split(".", 1)[1]
@@ -116,21 +54,20 @@ async def create_response(

        if request.stream:
            logger.info("Streaming response requested")
+
            if use_steward:
                logger.info("Streaming with Steward preprocessing for Tatlock request")
-                # Use Steward + Tatlock streaming (Milestone 3.5)
                from src.responses.streaming import StreamingCoordinator
                coordinator = StreamingCoordinator()
                return EventSourceResponse(
                    coordinator.stream_response_with_steward(request)
                )
            else:
-                # Regular streaming for non-Tatlock models
                return EventSourceResponse(
                    service.create_response_stream(request)
                )

-        # Use appropriate service method
+        # Non-streaming response
        if use_steward:
            logger.info("Using Steward preprocessing for Tatlock request")
            return await service.create_response_with_steward(request)
@@ -148,8 +85,3 @@ async def create_response(
    except Exception as e:
        logger.error(f"Unexpected error: {e}", exc_info=True)
        raise HTTPException(status_code=500, detail="Internal server error")
-
-    finally:
-        # Reset context (important for connection reuse)
-        current_user.reset(user_token)
-        current_conversation.reset(conv_token)
@@ -26,6 +26,8 @@ from src.responses.context import ContextWindow
 from src.core.preprocessing import preprocess_request
 from src.core.tool_tracking import ToolCallTracker
 from src.core.logging_config import get_logger
+from src.core.tracing import start_trace, end_trace, start_span, SpanType
+from src.core.context import current_user, current_conversation, get_default_user
 from src.agents.steward.schemas import StewardRecommendation

 import re
@@ -34,6 +36,29 @@ import asyncio
 logger = get_logger(__name__)


+def _extract_user_input(input_data) -> str:
+    """Extract user input text from request input for tracing."""
+    if isinstance(input_data, str):
+        return input_data
+    elif isinstance(input_data, list) and input_data:
+        last_msg = input_data[-1]
+        if isinstance(last_msg, dict):
+            return last_msg.get("content", str(last_msg))
+        return str(last_msg)
+    return ""
+
+
+def _extract_response_preview(response: Response) -> str:
+    """Extract response preview text for tracing."""
+    if response.output:
+        for item in response.output:
+            if hasattr(item, 'content'):
+                for content in item.content:
+                    if hasattr(content, 'text'):
+                        return content.text[:200]
+    return ""
+
+
 async def _execute_single_delegation(
    agent_name: str,
    task: str,
@@ -405,45 +430,88 @@ async def create_response(request: ResponseRequest) -> Response:
    # Get or generate conversation ID
    conversation_id = await _conversation_history.get_conversation_id(request)

-    # Strip pipeline prefix if present (e.g., "pipeline.model" -> "model")
-    model_id = request.model
-    if "." in model_id:
-        model_id = model_id.split(".", 1)[1]
+    # Set context for tracing
+    effective_user = request.user or get_default_user()
+    current_user.set(effective_user)
+    current_conversation.set(conversation_id)

-    # Get agent for model
-    agent = ModelRegistry.get_agent(model_id)
+    # Extract user input for tracing
+    user_input = _extract_user_input(request.input)

-    # Collect all output items from agent
-    output_items = []
-    async for item in agent.generate_response(
-        messages=request.input,
-        reasoning=request.reasoning,
-        tools=request.tools,
-        temperature=request.temperature,
-        max_tokens=request.max_output_tokens,
-        stop=request.stop,
-    ):
-        output_items.append(item)
-
-    # Convert agent OutputItems to schema OutputItems
-    converted_items = _convert_output_items(output_items)
-
-    # Calculate token usage
-    usage = _calculate_usage(request.input, output_items)
-
-    response = Response(
-        id=f"resp_{generate_id()}",
-        created_at=int(time.time()),
-        model=request.model,
-        status="completed",
-        output=converted_items,
-        usage=usage
+    # Start trace
+    trace = start_trace(
+        conversation_id=conversation_id,
+        user=effective_user,
+        request={
+            "model": request.model,
+            "input_preview": user_input[:200] if user_input else "",
+            "full_input": request.input,
+            "streaming": False,
+        },
    )

-    # Track conversation history (for analytics and future vector memory)
-    await _conversation_history.add_response(conversation_id, response)
+    # Start service span
+    service_span = start_span(
+        "create_response",
+        SpanType.ROUTER,
+        metadata={"model": request.model, "user": effective_user},
+    )

-    return response
+    try:
+        # Strip pipeline prefix if present (e.g., "pipeline.model" -> "model")
+        model_id = request.model
+        if "." in model_id:
+            model_id = model_id.split(".", 1)[1]
+
+        # Get agent for model
+        agent = ModelRegistry.get_agent(model_id)
+
+        # Collect all output items from agent
+        output_items = []
+        async for item in agent.generate_response(
+            messages=request.input,
+            reasoning=request.reasoning,
+            tools=request.tools,
+            temperature=request.temperature,
+            max_tokens=request.max_output_tokens,
+            stop=request.stop,
+        ):
+            output_items.append(item)
+
+        # Convert agent OutputItems to schema OutputItems
+        converted_items = _convert_output_items(output_items)
+
+        # Calculate token usage
+        usage = _calculate_usage(request.input, output_items)
+
+        response = Response(
+            id=f"resp_{generate_id()}",
+            created_at=int(time.time()),
+            model=request.model,
+            status="completed",
+            output=converted_items,
+            usage=usage
+        )
+
+        # Track conversation history (for analytics and future vector memory)
+        await _conversation_history.add_response(conversation_id, response)
+
+        # End trace with response info
+        response_preview = _extract_response_preview(response)
+        end_trace(
+            response={
+                "output_preview": response_preview,
+                "output_count": len(response.output) if response.output else 0,
+                "status": response.status,
+            },
+            status="completed",
+        )
+
+        return response
+
+    except Exception as e:
+        end_trace(status="error")
+        raise


 async def create_response_with_steward(request: ResponseRequest) -> Response:
@@ -454,7 +522,7 @@ async def create_response_with_steward(request: ResponseRequest) -> Response:
    1. Steward analyzes the request and recommends capabilities
    2. Phase 1: Tatlock orchestrates tool calls and expert delegations
    3. Phase 2: Tatlock synthesizes butler-toned response from results
-    4. Tool usage is tracked for benchmarking
+    4. Tool usage is tracked for analysis

    Args:
        request: Response request
@@ -473,133 +541,176 @@ async def create_response_with_steward(request: ResponseRequest) -> Response:
    # Get or generate conversation ID
    conversation_id = await _conversation_history.get_conversation_id(request)

-    # Extract user message and conversation history
-    user_message = ""
-    for msg in reversed(request.input):
-        if msg.get("role") == "user":
-            user_message = msg.get("content", "")
-            break
+    # Set context for tracing
+    effective_user = request.user or get_default_user()
+    current_user.set(effective_user)
+    current_conversation.set(conversation_id)

-    # Conversation history is all messages except the current one
-    conversation_history = request.input[:-1] if len(request.input) > 1 else []
+    # Extract user input for tracing
+    user_input = _extract_user_input(request.input)

-    logger.info(
-        "creating_response_with_steward",
-        user_message_preview=user_message[:100],
-        history_length=len(conversation_history),
+    # Start trace
+    trace = start_trace(
        conversation_id=conversation_id,
+        user=effective_user,
+        request={
+            "model": request.model,
+            "input_preview": user_input[:200] if user_input else "",
+            "full_input": request.input,
+            "streaming": False,
+        },
    )

-    # Steward preprocessing
-    enriched = await preprocess_request(
-        user_message,
-        conversation_history=conversation_history,
-        conversation_id=conversation_id,
+    # Start service span
+    service_span = start_span(
+        "create_response_with_steward",
+        SpanType.ROUTER,
+        metadata={"model": request.model, "user": effective_user},
    )

-    # Initialize tool tracker
-    tracker = ToolCallTracker(
-        recommended_capabilities=enriched.recommendation.recommended_capabilities,
-        conversation_id=conversation_id,
-    )
+    try:
+        # Extract user message and conversation history
+        user_message = ""
+        for msg in reversed(request.input):
+            if msg.get("role") == "user":
+                user_message = msg.get("content", "")
+                break

-    # Check if direct delegation is recommended
-    # If Steward recommends ONLY delegation agents (biographer/librarian/housekeeper),
-    # we still use two-phase but delegate directly in Phase 1
-    delegation_agents = {"biographer", "librarian", "housekeeper"}
-    delegation_only = all(
-        cap in delegation_agents
-        for cap in enriched.recommendation.recommended_capabilities
-    ) and enriched.recommendation.recommended_capabilities
+        # Conversation history is all messages except the current one
+        conversation_history = request.input[:-1] if len(request.input) > 1 else []

-    from src.agents.tatlock import TatlockAgent
-    tatlock = TatlockAgent()
-
-    # Use enriched query (with location/timezone context) if available
-    effective_query = enriched.recommendation.enriched_query or user_message
-
-    if delegation_only:
-        # Direct delegation path - collect results then synthesize
-        orchestration_results = await _direct_delegation_with_results(
-            effective_query, enriched.recommendation, tracker, conversation_id
-        )
-    else:
-        # Phase 1: Orchestrate tool calls
-        orchestration_results = await tatlock.orchestrate_tool_calls(
-            user_message=effective_query,
-            steward_note=enriched.steward_note,
-            scoped_tools=enriched.scoped_tools,
-            message_history=conversation_history,
-            tool_tracker=tracker,
+        logger.info(
+            "creating_response_with_steward",
+            user_message_preview=user_message[:100],
+            history_length=len(conversation_history),
+            conversation_id=conversation_id,
        )

-        # Handle text-based delegation fallback if present
-        if "[DELEGATE:" in orchestration_results.get("raw_output", ""):
-            text_delegation_results = await _handle_text_delegation(
-                orchestration_results["raw_output"], tracker, conversation_id
+        # Steward preprocessing
+        enriched = await preprocess_request(
+            user_message,
+            conversation_history=conversation_history,
+            conversation_id=conversation_id,
+        )
+
+        # Initialize tool tracker
+        tracker = ToolCallTracker(
+            recommended_capabilities=enriched.recommendation.recommended_capabilities,
+            conversation_id=conversation_id,
+        )
+
+        # Check if direct delegation is recommended
+        # If Steward recommends ONLY delegation agents (biographer/librarian/housekeeper),
+        # we still use two-phase but delegate directly in Phase 1
+        delegation_agents = {"biographer", "librarian", "housekeeper"}
+        delegation_only = all(
+            cap in delegation_agents
+            for cap in enriched.recommendation.recommended_capabilities
+        ) and enriched.recommendation.recommended_capabilities
+
+        from src.agents.tatlock import TatlockAgent
+        tatlock = TatlockAgent()
+
+        # Use enriched query (with location/timezone context) if available
+        effective_query = enriched.recommendation.enriched_query or user_message
+
+        if delegation_only:
+            # Direct delegation path - collect results then synthesize
+            orchestration_results = await _direct_delegation_with_results(
+                effective_query, enriched.recommendation, tracker, conversation_id
+            )
+        else:
+            # Phase 1: Orchestrate tool calls
+            orchestration_results = await tatlock.orchestrate_tool_calls(
+                user_message=effective_query,
+                steward_note=enriched.steward_note,
+                scoped_tools=enriched.scoped_tools,
+                message_history=conversation_history,
+                tool_tracker=tracker,
            )
-            # Add text delegation results to expert_results
-            if text_delegation_results != orchestration_results["raw_output"]:
-                orchestration_results["expert_results"]["text_delegation"] = text_delegation_results

-    # Phase 2: Synthesize butler-toned response from all results
-    tatlock_response = await tatlock.synthesize_from_results(
-        user_message=user_message,
-        orchestration_results=orchestration_results,
-        message_history=conversation_history,
-    )
+            # Handle text-based delegation fallback if present
+            if "[DELEGATE:" in orchestration_results.get("raw_output", ""):
+                text_delegation_results = await _handle_text_delegation(
+                    orchestration_results["raw_output"], tracker, conversation_id
+                )
+                # Add text delegation results to expert_results
+                if text_delegation_results != orchestration_results["raw_output"]:
+                    orchestration_results["expert_results"]["text_delegation"] = text_delegation_results

-    # Finalize tool tracking
-    await tracker.finalize()
+        # Phase 2: Synthesize butler-toned response from all results
+        tatlock_response = await tatlock.synthesize_from_results(
+            user_message=user_message,
+            orchestration_results=orchestration_results,
+            message_history=conversation_history,
+        )

-    # Build response output items
-    output_items = []
+        # Finalize tool tracking
+        await tracker.finalize()

-    # Add Steward reasoning as a reasoning output item
-    output_items.append(ReasoningOutputItem(
-        id=f"reasoning_{generate_id()}",
-        summary=[
-            "🎩 Steward's Analysis:",
-            enriched.steward_reasoning,
-        ],
-        status="completed"
-    ))
+        # Build response output items
+        output_items = []

-    # Add Tatlock's message
-    output_items.append(MessageOutputItem(
-        id=f"msg_{generate_id()}",
-        role="assistant",
-        content=[OutputTextContent(
-            type="output_text",
-            text=tatlock_response,
-            annotations=[]
-        )],
-        status="completed"
-    ))
+        # Add Steward reasoning as a reasoning output item
+        output_items.append(ReasoningOutputItem(
+            id=f"reasoning_{generate_id()}",
+            summary=[
+                "🎩 Steward's Analysis:",
+                enriched.steward_reasoning,
+            ],
+            status="completed"
+        ))

-    # Calculate usage (approximate)
-    usage = _calculate_usage(request.input, output_items)
+        # Add Tatlock's message
+        output_items.append(MessageOutputItem(
+            id=f"msg_{generate_id()}",
+            role="assistant",
+            content=[OutputTextContent(
+                type="output_text",
+                text=tatlock_response,
+                annotations=[]
+            )],
+            status="completed"
+        ))

-    response = Response(
-        id=f"resp_{generate_id()}",
-        created_at=int(time.time()),
-        model=request.model,
-        status="completed",
-        output=output_items,
-        usage=usage
-    )
+        # Calculate usage (approximate)
+        usage = _calculate_usage(request.input, output_items)

-    # Track conversation history
-    await _conversation_history.add_response(conversation_id, response)
+        response = Response(
+            id=f"resp_{generate_id()}",
+            created_at=int(time.time()),
+            model=request.model,
+            status="completed",
+            output=output_items,
+            usage=usage
+        )

-    logger.info(
-        "response_with_steward_complete",
-        response_id=response.id,
-        recommended_capabilities=enriched.recommendation.recommended_capabilities,
-        tool_summary=tracker.get_summary(),
-    )
+        # Track conversation history
+        await _conversation_history.add_response(conversation_id, response)

-    return response
+        logger.info(
+            "response_with_steward_complete",
+            response_id=response.id,
+            recommended_capabilities=enriched.recommendation.recommended_capabilities,
+            tool_summary=tracker.get_summary(),
+        )
+
+        # End trace with response info
+        response_preview = _extract_response_preview(response)
+        end_trace(
+            response={
+                "output_preview": response_preview,
+                "output_count": len(response.output) if response.output else 0,
+                "status": response.status,
+            },
+            status="completed",
+        )
+
+        return response
+
+    except Exception as e:
+        end_trace(status="error")
+        raise


 async def create_response_stream(
@@ -1,352 +0,0 @@
-"""
-Tests for benchmark storage.
-
-Tests performance tracking, Redis storage, and analytics features.
-"""
-import json
-from datetime import datetime, timedelta, timezone
-from unittest.mock import AsyncMock, MagicMock, patch
-
-import pytest
-
-from src.core.benchmarks import (
-    BenchmarkStore,
-    PerformanceBenchmark,
-    get_benchmark_store,
-)
-
-
-class TestPerformanceBenchmark:
-    """Test PerformanceBenchmark model."""
-
-    def test_benchmark_creation(self):
-        """Test creating a performance benchmark."""
-        benchmark = PerformanceBenchmark(
-            operation="steward_analysis",
-            duration_seconds=1.23,
-            success=True,
-            recommendation_count=3,
-        )
-
-        assert benchmark.operation == "steward_analysis"
-        assert benchmark.duration_seconds == 1.23
-        assert benchmark.success is True
-        assert benchmark.recommendation_count == 3
-        assert isinstance(benchmark.timestamp, datetime)
-
-    def test_benchmark_with_tool_fields(self):
-        """Test benchmark with tool-specific fields."""
-        benchmark = PerformanceBenchmark(
-            operation="tool_call",
-            duration_seconds=0.5,
-            success=True,
-            tool_name="calculate",
-            was_recommended=True,
-            was_actually_used=True,
-        )
-
-        assert benchmark.tool_name == "calculate"
-        assert benchmark.was_recommended is True
-        assert benchmark.was_actually_used is True
-
-    def test_benchmark_to_redis_dict(self):
-        """Test conversion to Redis dict."""
-        benchmark = PerformanceBenchmark(
-            operation="test_op",
-            duration_seconds=1.0,
-            success=True,
-            metadata={"key": "value"},
-        )
-
-        redis_dict = benchmark.to_redis_dict()
-        assert redis_dict["operation"] == "test_op"
-        assert redis_dict["duration_seconds"] == 1.0
-        assert redis_dict["success"] == "True"  # Booleans stored as strings in Redis
-        assert isinstance(redis_dict["timestamp"], str)
-        assert isinstance(redis_dict["metadata"], str)
-
-    def test_benchmark_from_redis_dict(self):
-        """Test reconstruction from Redis dict."""
-        now = datetime.now(timezone.utc)
-        redis_dict = {
-            "timestamp": now.isoformat(),
-            "operation": "test_op",
-            "duration_seconds": 1.5,
-            "success": "True",  # Booleans stored as strings in Redis
-            "metadata": json.dumps({"test": "data"}),
-            "recommendation_count": None,
-            "confidence": None,
-            "tool_name": None,
-            "was_recommended": None,
-            "was_actually_used": None,
-            "conversation_id": None,
-        }
-
-        benchmark = PerformanceBenchmark.from_redis_dict(redis_dict)
-        assert benchmark.operation == "test_op"
-        assert benchmark.duration_seconds == 1.5
-        assert benchmark.success is True  # Converted back to bool
-        assert benchmark.metadata == {"test": "data"}
-
-
-class TestBenchmarkStore:
-    """Test BenchmarkStore functionality."""
-
-    @pytest.fixture
-    def mock_redis(self):
-        """Create mock Redis client."""
-        mock = AsyncMock()
-        mock.hset = AsyncMock()
-        mock.expire = AsyncMock()
-        mock.zadd = AsyncMock()
-        mock.zrevrangebyscore = AsyncMock(return_value=[])
-        mock.hgetall = AsyncMock(return_value={})
-        mock.aclose = AsyncMock()
-        return mock
-
-    @pytest.fixture
-    def store(self, mock_redis):
-        """Create benchmark store with mock Redis."""
-        return BenchmarkStore(redis_client=mock_redis)
-
-    @pytest.mark.asyncio
-    async def test_record_benchmark(self, store, mock_redis):
-        """Test recording a benchmark."""
-        benchmark = PerformanceBenchmark(
-            operation="test_op",
-            duration_seconds=1.0,
-            success=True,
-        )
-
-        await store.record(benchmark)
-
-        # Verify Redis calls
-        mock_redis.hset.assert_called_once()
-        mock_redis.expire.assert_called()
-        mock_redis.zadd.assert_called_once()
-
-    @pytest.mark.asyncio
-    async def test_record_benchmark_disabled(self, mock_redis):
-        """Test recording when benchmarks are disabled."""
-        with patch("src.core.benchmarks.config.ENABLE_BENCHMARKS", False):
-            store = BenchmarkStore(redis_client=mock_redis)
-            benchmark = PerformanceBenchmark(
-                operation="test_op",
-                duration_seconds=1.0,
-                success=True,
-            )
-
-            await store.record(benchmark)
-
-            # Should not call Redis
-            mock_redis.hset.assert_not_called()
-
-    @pytest.mark.asyncio
-    async def test_record_benchmark_handles_errors(self, store, mock_redis):
-        """Test recording handles Redis errors gracefully."""
-        mock_redis.hset.side_effect = Exception("Redis error")
-
-        benchmark = PerformanceBenchmark(
-            operation="test_op",
-            duration_seconds=1.0,
-            success=True,
-        )
-
-        # Should not raise exception
-        await store.record(benchmark)
-
-    @pytest.mark.asyncio
-    async def test_query_benchmarks(self, store, mock_redis):
-        """Test querying benchmarks."""
-        # Setup mock data
-        now = datetime.now(timezone.utc)
-        mock_key = f"benchmark:test_op:{int(now.timestamp() * 1000)}"
-        mock_redis.zrevrangebyscore.return_value = [mock_key]
-
-        # Mock hgetall to return proper data (booleans as strings, like Redis)
-        mock_redis.hgetall.return_value = {
-            "timestamp": now.isoformat(),
-            "operation": "test_op",
-            "duration_seconds": 1.5,  # Numeric, not string
-            "success": "True",  # Booleans stored as strings in Redis
-            "metadata": "{}",
-            "recommendation_count": None,
-            "confidence": None,
-            "tool_name": None,
-            "was_recommended": None,
-            "was_actually_used": None,
-            "conversation_id": None,
-        }
-
-        results = await store.query("test_op", limit=10)
-
-        assert len(results) == 1
-        assert results[0].operation == "test_op"
-        mock_redis.zrevrangebyscore.assert_called_once()
-
-    @pytest.mark.asyncio
-    async def test_query_with_time_range(self, store, mock_redis):
-        """Test querying with time range."""
-        now = datetime.now(timezone.utc)
-        start_time = now - timedelta(hours=1)
-        end_time = now
-
-        await store.query("test_op", start_time=start_time, end_time=end_time)
-
-        # Verify time range was converted to timestamps
-        call_args = mock_redis.zrevrangebyscore.call_args
-        assert call_args is not None
-
-    @pytest.mark.asyncio
-    async def test_query_disabled_benchmarks(self, mock_redis):
-        """Test querying when benchmarks are disabled."""
-        with patch("src.core.benchmarks.config.ENABLE_BENCHMARKS", False):
-            store = BenchmarkStore(redis_client=mock_redis)
-            results = await store.query("test_op")
-            assert results == []
-
-    @pytest.mark.asyncio
-    async def test_query_handles_errors(self, store, mock_redis):
-        """Test query handles errors gracefully."""
-        mock_redis.zrevrangebyscore.side_effect = Exception("Redis error")
-
-        results = await store.query("test_op")
-        assert results == []
-
-    @pytest.mark.asyncio
-    async def test_get_statistics(self, store, mock_redis):
-        """Test getting statistics."""
-        # Setup mock data with multiple benchmarks
-        now = datetime.now(timezone.utc)
-        mock_keys = [
-            f"benchmark:test_op:{int((now - timedelta(seconds=i)).timestamp() * 1000)}"
-            for i in range(3)
-        ]
-        mock_redis.zrevrangebyscore.return_value = mock_keys
-
-        # Return different durations and success values
-        benchmarks_data = [
-            {"duration_seconds": "1.0", "success": "True"},
-            {"duration_seconds": "2.0", "success": "True"},
-            {"duration_seconds": "3.0", "success": "False"},
-        ]
-
-        async def mock_hgetall(key):
-            idx = mock_keys.index(key)
-            data = benchmarks_data[idx]
-            return {
-                "timestamp": now.isoformat(),
-                "operation": "test_op",
-                "duration_seconds": float(data["duration_seconds"]),
-                "success": data["success"],  # Pass string through, from_redis_dict converts
-                "metadata": "{}",
-                "recommendation_count": None,
-                "confidence": None,
-                "tool_name": None,
-                "was_recommended": None,
-                "was_actually_used": None,
-                "conversation_id": None,
-            }
-
-        mock_redis.hgetall.side_effect = mock_hgetall
-
-        stats = await store.get_statistics("test_op")
-
-        assert stats["count"] == 3
-        assert stats["avg_duration"] == 2.0  # (1 + 2 + 3) / 3
-        assert stats["min_duration"] == 1.0
-        assert stats["max_duration"] == 3.0
-        assert stats["success_rate"] == pytest.approx(66.67, rel=0.01)
-        assert stats["total_successes"] == 2
-        assert stats["total_failures"] == 1
-
-    @pytest.mark.asyncio
-    async def test_get_statistics_empty(self, store, mock_redis):
-        """Test statistics with no data."""
-        mock_redis.zrevrangebyscore.return_value = []
-
-        stats = await store.get_statistics("test_op")
-
-        assert stats["count"] == 0
-        assert stats["avg_duration"] == 0.0
-        assert stats["success_rate"] == 0.0
-
-    @pytest.mark.asyncio
-    async def test_get_tool_accuracy(self, store, mock_redis):
-        """Test tool accuracy calculation."""
-        # Setup mock data
-        now = datetime.now(timezone.utc)
-        mock_keys = [
-            f"benchmark:tool_call:{int((now - timedelta(seconds=i)).timestamp() * 1000)}"
-            for i in range(4)
-        ]
-        mock_redis.zrevrangebyscore.return_value = mock_keys
-
-        # Different combinations of recommended/used
-        tool_data = [
-            {"was_recommended": "True", "was_actually_used": "True"},  # Good
-            {"was_recommended": "True", "was_actually_used": "True"},  # Good
-            {"was_recommended": "False", "was_actually_used": "True"},  # Missed
-            {"was_recommended": "True", "was_actually_used": "False"},  # Not used
-        ]
-
-        async def mock_hgetall(key):
-            idx = mock_keys.index(key)
-            data = tool_data[idx]
-            return {
-                "timestamp": now.isoformat(),
-                "operation": "tool_call",
-                "duration_seconds": 1.0,
-                "success": "True",  # Booleans stored as strings in Redis
-                "metadata": "{}",
-                "recommendation_count": None,
-                "confidence": None,
-                "tool_name": "test_tool",
-                "conversation_id": None,
-                "was_recommended": data["was_recommended"],  # Already strings
-                "was_actually_used": data["was_actually_used"],  # Already strings
-            }
-
-        mock_redis.hgetall.side_effect = mock_hgetall
-
-        accuracy = await store.get_tool_accuracy()
-
-        assert accuracy["total_calls"] == 4
-        assert accuracy["total_used"] == 3
-        assert accuracy["recommended_and_used"] == 2
-        assert accuracy["not_recommended_but_used"] == 1
-        assert accuracy["precision"] == pytest.approx(66.67, rel=0.01)
-
-    @pytest.mark.asyncio
-    async def test_get_tool_accuracy_empty(self, store, mock_redis):
-        """Test tool accuracy with no data."""
-        mock_redis.zrevrangebyscore.return_value = []
-
-        accuracy = await store.get_tool_accuracy()
-
-        assert accuracy["total_calls"] == 0
-        assert accuracy["precision"] == 0.0
-
-    @pytest.mark.asyncio
-    async def test_close(self, store, mock_redis):
-        """Test closing the store."""
-        await store.close()
-        mock_redis.aclose.assert_called_once()
-
-        # Client should be None after close
-        assert store._client is None
-
-
-class TestGlobalBenchmarkStore:
-    """Test global benchmark store instance."""
-
-    def test_get_benchmark_store(self):
-        """Test getting global store instance."""
-        store = get_benchmark_store()
-        assert isinstance(store, BenchmarkStore)
-
-    def test_get_benchmark_store_singleton(self):
-        """Test store is singleton."""
-        store1 = get_benchmark_store()
-        store2 = get_benchmark_store()
-        assert store1 is store2
@@ -43,8 +43,8 @@ LOG_FILE="$LOGS_DIR/server.log"
 echo -e "${YELLOW}Logs will be written to: ${LOG_FILE}${NC}"

 # Start the server
-echo -e "${GREEN}Starting uvicorn server on http://localhost:8777${NC}"
+echo -e "${GREEN}Starting uvicorn server on http://tower-of-joy:8777${NC}"
 echo -e "${YELLOW}Press Ctrl+C to stop the server${NC}"
 echo ""

-uvicorn src.main:app --reload --host localhost --port 8777 2>&1 | tee "$LOG_FILE"
+uvicorn src.main:app --reload --host 0.0.0.0 --port 8777 2>&1 | tee "$LOG_FILE"
Author	SHA1	Message	Date
jpmschweitzerandClaude Opus 4.5	496f37a538	feat: add Claude backend with automatic Ollama fallback (Claudification Phase 1) Build and Push / release (release) Failing after 6s Details Build and Push / build (release) Successful in 3m5s Details All agents now prefer Claude API when ANTHROPIC_API_KEY is configured, with automatic fallback to Ollama when offline or unconfigured. New src/anthropic/ module provides model selection via get_model() factory. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2026-02-05 07:19:29 +01:00
jpmschweitzer	5d23bcae79	auto release/build on version tag	2026-01-03 20:39:51 +01:00
jpmschweitzerandClaude Opus 4.5	62eac3eb61	feat: integrate Paperless documents and volatile cache into Librarian Build and Push / build (release) Successful in 54s Details - Add Paperless document search to HybridRAG pipeline - Add volatile cache (weather, forecast, news, stocks) to HybridRAG - Add include_documents and include_volatile params to hybrid_search - Add 📑 and ⚡ icons for document/volatile sources - Update Librarian prompt with new data source awareness - Fix Biographer routing: personal memory queries now route correctly - Add location keywords to Steward pre-fetch logic 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2025-12-30 13:31:23 +01:00
jpmschweitzerandClaude Opus 4.5	ba195e401a	fix: reduce Tatlock's excessive apologizing Build and Push / build (release) Successful in 53s Details Strengthened personality prompt to prevent unnecessary apologies after successful Librarian delegations. Added explicit "do NOT apologize" instructions to both system prompt and synthesis prompt. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2025-12-23 17:52:19 +01:00
jpmschweitzerandClaude Opus 4.5	3ec4f402fa	chore: release v1.10.0 Build and Push / build (release) Successful in 1m2s Details 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2025-12-22 10:34:12 +01:00
jpmschweitzerandClaude Opus 4.5	628f05532b	chore: bind server to all network interfaces Change uvicorn from localhost to 0.0.0.0 to allow connections from other machines on the network. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2025-12-22 10:28:24 +01:00
jpmschweitzerandClaude Opus 4.5	51fd59ce92	fix: prevent Librarian from fabricating information Add explicit instructions to the Librarian system prompt to never invent data when tools fail or data sources are unavailable. - Report what failed specifically - Never provide placeholder or made-up data - Better to return no information than fabricated information 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2025-12-22 10:28:13 +01:00
jpmschweitzerandClaude Opus 4.5	87f2926db2	feat: integrate tracing throughout request pipeline Instrument the full request flow with trace spans for debugging: - Wrap expert delegations (librarian/biographer/housekeeper) in spans - Add orchestrate and synthesize spans to TatlockAgent - Trace Steward analysis in preprocessing - Start/end traces in response service with context management - Simplify router by moving context handling to service layer - Include tracing router in debug mode - Remove benchmark recording from tool_tracking and steward service 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2025-12-22 10:26:37 +01:00
jpmschweitzerandClaude Opus 4.5	2a9449bc81	refactor: remove Redis benchmark system Remove the Redis-backed performance benchmarking in favor of the new lightweight file-based tracing system which provides better debugging capabilities for local development. - Delete src/core/benchmarks.py - Remove ENABLE_BENCHMARKS, REDIS_BENCHMARK_DB, redis_url from config - Update memory_cache comment (now uses DB 1) 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2025-12-22 10:26:17 +01:00
jpmschweitzerandClaude Opus 4.5	60d84535c0	feat: add lightweight request tracing for debugging Adds JSON-based tracing system for local development that captures the full request flow through Tatlock's multi-agent architecture. - Trace/Span dataclasses with automatic timing and nesting - Context-var based propagation for async-safe tracing - trace_span async context manager for clean instrumentation - Traces written to logs/traces/{trace_id}.json - REST API for listing and retrieving traces (/traces) - Standalone HTML viewer with timeline visualization Enabled via DEBUG=true environment variable. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>	2025-12-22 10:25:56 +01:00