ypollak2/llm-router

v8.2.0 Feature

This release adds 3 notable features for engineering teams evaluating rollout.

Published 21d LLM Frameworks

View tool

✓ 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

ReleasePort's take

Light signal

editorial:auto 13d

v8.2.0 adds always-on routing explainability showing model selection rationale and cost savings across four configurable modes. Includes Ollama fallback for RESEARCH chains and 27 new tests for improved reliability.

Why it matters: Operators now see why each model was chosen and savings achieved without extra configuration. Verbose mode displays full chain diagnostics. Ollama fallback prevents complete RESEARCH chain failure. Test coverage improvement strengthens production reliability.

Summary

AI summary

Routing rationale footer now shows why a model was chosen and savings

Changes in this release

Type	Severity	Summary	CVE
Feature
Feature	Medium	Every routed response now shows why it was routed and savings amount. Every routed response now shows why it was routed and savings amount. Source: llm_adapter@2026-05-21 Confidence: high	—
Feature	Medium	Always-on routing rationale footer added without requiring env var. Always-on routing rationale footer added without requiring env var. Source: llm_adapter@2026-05-21 Confidence: high	—
Feature	Medium	Four explainability modes: footer (default), header, verbose, off. Four explainability modes: footer (default), header, verbose, off. Source: llm_adapter@2026-05-21 Confidence: high	—
Feature	Medium	Verbose mode displays full chain walk with per-model attempt status. Verbose mode displays full chain walk with per-model attempt status. Source: llm_adapter@2026-05-21 Confidence: high	—
Feature	Medium	Savings calculation added independently, no database dependency. Savings calculation added independently, no database dependency. Source: llm_adapter@2026-05-21 Confidence: high	—
Feature	Medium	Ollama fallback enabled in all RESEARCH chains; research never fails completely. Ollama fallback enabled in all RESEARCH chains; research never fails completely. Source: llm_adapter@2026-05-21 Confidence: high	—
Feature	Medium	Full chain error diagnostics added; view every model tried and failure reasons. Full chain error diagnostics added; view every model tried and failure reasons. Source: llm_adapter@2026-05-21 Confidence: high	—
Bugfix	Medium	Added 27 new tests to improve reliability. Added 27 new tests to improve reliability. Source: llm_adapter@2026-05-21 Confidence: low	—

Full changelog

What's New

Every routed response now shows why it was routed and how much you saved:

→ gemini-2.5-flash · simple · $0.0002 (43x cheaper)

Added

Always-on routing rationale footer (no env var needed)
4 explainability modes: footer (default), header, verbose, off
verbose mode shows full chain walk with per-model attempt status
Savings calculation (independent, no DB dependency)
Ollama fallback in all RESEARCH chains — research never fails completely
Full chain error diagnostics — see every model tried and why it failed
27 new tests

Configuration

export LLM_ROUTER_EXPLAIN=footer   # default — compact one-line after response
export LLM_ROUTER_EXPLAIN=verbose  # full breakdown with chain walk
export LLM_ROUTER_EXPLAIN=off      # disable

Upgrade

pip install --upgrade llm-routing

Full Changelog: https://github.com/ypollak2/llm-router/compare/v8.1.0...v8.2.0

View diff on GitHub

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

Share on X Share on Bluesky

Track ypollak2/llm-router

Get notified when new releases ship.

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

Related tools

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.