This release adds 4 notable features for engineering teams evaluating rollout.
✓ No known CVEs patched in this version
Topics
+7 more
Summary
AI summaryFour automated loops add self‑tuning routing and a community profile fallback, plus a critical bug fix for model recording.
Full changelog
What's new: Closed Loops Integration
🔄 Loop 1: Directives Feedback
- Load and apply ROUTING_RULE directives from persistent memory
- Session-start displays active directives
- Auto-route applies directives as classifier overrides
📊 Loop 2: Weekly Retrospective
- Aggregate patterns from 7 days of retrospectives
- Auto-generate ROUTING_RULE directives based on recurring patterns
- Callable via
llm-router retrospect --weekly
📈 Loop 3: Trend Pressure → Routing
- Monitor accuracy trends from hourly snapshots
- Quality decline triggers soft model escalation (independent of budget)
- Blends with budget pressure for intelligent model selection
🌍 Loop 4: Community Profile Default
- Fallback community URL for one-click profile imports
- Enables discovery of community-maintained routing patterns
🐛 Critical Bug Fix
- Fixed:
llm_reroutenow properly records corrected_model to database - Learned profiles now capture actual models (no more model="unknown")
Architecture
All 4 loops close one-way data flows from v6.1:
- Corrections → learned profiles → directives ✓
- Retrospectives → actions → directives ✓
- Snapshots → trends → model selection ✓
- Profile imports → community sharing ✓
Upgrade
```bash
pip install --upgrade claude-code-llm-router && llm-router install
```
Testing
- 13 new closed-loops tests
- 21 model selector tests (pressure blending verified)
- All 34 tests passing
- No regressions in full suite
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About ypollak2/llm-router
Subscription-aware LLM router for Claude Code. Routes tasks to 20+ providers (OpenAI, Gemini, Groq, Ollama, Codex) based on complexity classification, Claude subscription pressure, and cost. Free tasks stay on Claude subscription; expensive tasks fall back to the cheapest capable model. Includes 30 MCP tools, 6 auto-routing hooks, semantic dedup cache, prompt caching, daily spend cap, and a live web dashboard.
Related context
Related tools
Beta — feedback welcome: [email protected]