Skip to content

ypollak2/llm-router

v7.4.0 Breaking

This release includes breaking changes for platform teams planning a safe upgrade.

Published 1mo LLM Frameworks
✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

ai-routing anthropic claude claude-code cost-optimization gemini
+7 more
litellm llm llm-router mcp-server model-router ollama openai

Summary

AI summary

Automatic content‑generation detection routes write/draft patterns via llm_generate to achieve ~90% cost savings.

Full changelog

Content Generation Routing Discipline

Systemic routing improvements to prevent content generation routing misses across all llm-router users.

What's New

  • Automatic Content Generation Detection — Hook detects "write", "draft", "compose", "add card", "add section" patterns
  • Fast-Path Routing — Detected patterns route instantly via llm_generate before classifier layers
  • Decomposition Guidance — CLAUDE.md decision trees codify best practices: generate first, integrate locally second
  • Soft Nudge Suggestions — Non-blocking suggestions encourage routing discipline without enforcement

Cost Impact

  • Per task: ~$0.0009 saved (route $0.0001 vs local $0.001 generation) = 90% savings
  • Per cycle: $0.10–$0.30 saved across 20-30 content tasks per 51-release cycle
  • Per team: $0.36–$0.54 annual savings for 5-engineer team

How It Works

When you write something like "add carousel card about AI auditing to CAROUSEL-BLUEPRINT.md":

  1. Detect — Hook recognizes "add.*card" + ".md file" pattern
  2. Route — Suggests: "Consider routing via llm_generate first, then integrate locally"
  3. Execute — Users follow suggestion: llm_generate (cheap) → Edit (local integration)
  4. Learn — Repeated nudges form habit; users adopt routing discipline by default

Documentation

  • CLAUDE.md — New § Content Generation Routing (v7.4.0)

    • Pre-flight decision tree: "Does this task create new content?"
    • Cost impact quantified with examples
    • Decomposition pattern guidelines
  • README.md — New v7.4.0 features section

    • Highlights automatic detection + pattern matching
    • References routing decision rules
  • CHANGELOG.md — Full implementation details

    • Detection patterns and fast-path routing
    • Technical architecture and cost analysis

Testing

  • ✅ 24/24 auto-route hook tests passing
  • ✅ Pattern specificity verified (no false positives)
  • ✅ Backward-compatible (fully soft enforcement)

Breaking Changes

None — fully backward-compatible. Detection is opt-in via hook suggestion.


Install: pip install --upgrade claude-code-llm-router

See CHANGELOG.md for full details.

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

Track ypollak2/llm-router

Get notified when new releases ship.

Sign up free

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.

All releases →

Related context

Earlier breaking changes

  • v9.2.0 Changes auto‑route directive from advisory "DO NOT SKIP" to hard constraint with explicit blocked tools list.
  • v9.2.0 Breaks permanent downgrade of enforcement after first Edit/Write; v13 now requires per‑turn routing.

Beta — feedback welcome: [email protected]