NadirRouter/NadirClaw

Open-source LLM router & AI cost optimizer. Routes simple prompts to cheap/local models, complex ones to premium — automatically. Drop-in OpenAI-compatible proxy for Claude Code, Codex, Cursor, OpenClaw. Saves 40-70% on AI API costs. Self-hosted, no middleman.

What it solves

NadirClaw reduces LLM API costs by 40-70% by ensuring that prompts are routed to the cheapest possible model capable of providing a reliable answer. It prevents the waste of expensive premium models on simple tasks (like formatting JSON or basic questions) that can be handled by smaller, cheaper models.

How it works

NadirClaw acts as a local drop-in proxy that uses a three-step process for every request:

  1. Route: A fast embedding classifier (~10ms) predicts the smallest viable model for the prompt. It can be overridden by specific markers like vision content or agentic tool loops.
  2. Verify: The response from the cheap model is checked against quality heuristics (such as checking for refusals, truncation, or JSON failures).
  3. Escalate: If the response fails the verification threshold, NadirClaw automatically escalates the request to the next-best, more powerful model.

Who it’s for

It is designed for solo developers and self-hosters who use OpenAI-compatible tools (like Cursor, Aider, or Continue) and want to lower their API spend without sacrificing quality.

Highlights

  • Drop-in Proxy: Compatible with any OpenAI-compatible client and supports Anthropic-native endpoints.
  • Context Optimization: Compacts bloated context (JSON, whitespace, etc.) to save 30-70% of input tokens.
  • Local Execution: Runs locally so API keys never leave the user's machine.
  • Cost Tracking: Includes a built-in dashboard, budget alerts, and a savings calculator.
  • Multi-Provider Support: Works with Gemini, OpenAI, Anthropic, Ollama, and any provider supported by LiteLLM.

Related

  • Project
    Continuum-AI-Corp/OrcaRouter-Lite

    A self-hosted, OpenAI-compatible LLM router that automatically routes requests to the cheapest capable model across multiple providers.

  • Project
    workweave/router

    A drop-in LLM proxy that automatically routes every request to the best-performing model using a local embedder and cluster scoring.

  • Project
    jia-gao/leanctx

    A drop-in prompt compression library that reduces LLM input-token costs by 10-40% using loss-tolerance routing to protect critical code and tool calls while compressing prose.

  • Project
    ulab-uiuc/LLMRouter

    An open-source library for intelligent LLM routing that dynamically selects the optimal model for each query to balance cost and performance.

  • Dispatch
    Weave Router: Intelligent Model Routing for Agentic Systems

    Weave Router is a drop-in proxy that reduces LLM costs by 40-70% by routing prompts to the most efficient model in under 50ms using a cluster scorer derived from Avengers-Pro.