Claude Skill

NadirRouter/NadirClaw

NadirClaw is an open-source LLM router that automatically routes simple prompts to cheap/local models and complex ones to premium. Drop-in OpenAI proxy for Claude Code, Codex, Cursor, OpenClaw. Sel...

Overview

Stars536
Forks66
LanguagePython
Last pushed2026-06-12
Last synced2026-06-17
View on GitHub

Repository

OwnerNadirRouter
RepositoryNadirClaw
Full nameNadirRouter/NadirClaw
Repo ID1,155,005,282

Install this Skill

pip install nadirclaw

Registry

Typeopenclaw_skill
Quality score85/100
Verificationreadme_parsed
Last verified2026-06-09
Platforms
ClaudeOpenClawCodexCursor
Capabilities
browsermemorysearchimageterminalworkflowaiai-cost-reductionai-routerclaude-code
Detected files
README.mddocker-compose.ymldocspyproject.tomltests
Config keys
GEMINI_API_KEYOPENAI_API_KEYANTHROPIC_API_KEYGOOGLE_API_KEYDEEPSEEK_API_KEYURLANTHROPIC_BASE_URLNADIRCLAW_MODEL_REGISTRY_URLNADIRCLAW_AUTH_TOKENNADIRCLAW_BUDGET_WEBHOOK_URL
Install methods
  • pip install nadirclaw
  • git clone https://github.com/doramirdor/NadirClaw.git
  • pip install -e .
  • git clone https://github.com/doramirdor/NadirClaw.git && cd NadirClaw
  • docker run -p 8856:8856 --env-file .env nadirclaw

Summary

NadirClaw is an open-source LLM router and AI cost optimizer that automatically routes simple prompts to cheap or local models and complex ones to premium models. It acts as a drop-in OpenAI-compatible proxy for tools like Claude Code, Codex, Cursor, and OpenClaw, saving 40-70% on AI API costs. Self-hosted with no middleman.

Chinese description

开源LLM路由器与AI成本优化器。自动将简单提示路由至廉价/本地模型,复杂提示路由至高端模型。可作为Claude Code、Codex、Cursor、OpenClaw的即插即用型OpenAI兼容代理。可节省40-70%的AI API成本。支持自托管,无需中间商。

Key features

  • Automatic prompt routing based on complexity
  • Drop-in OpenAI-compatible proxy for popular AI tools
  • Self-hosted, no third-party middleman
  • Supports local models via Ollama and premium models like Gemini
  • Saves 40-70% on AI API costs
  • Written in Python, easy to deploy and extend

Use cases

  • Reduce API costs for Claude Code and Codex workflows
  • Optimize AI spending in Cursor IDE
  • Route simple queries to local models via Ollama
  • Deploy a cost-aware proxy for OpenClaw
  • Self-host a smart LLM gateway for development teams

README excerpt

<p align="center"> <a href="https://getnadir.com"> <img src="docs/images/banner.png" alt="NadirClaw — Cut LLM & Agent Costs 40-70%" width="100%" /> </a> </p> <h1 align="center">NadirClaw</h1> <p align="center"> <strong>Your simple prompts are burning premium tokens.</strong><br> NadirClaw routes them to cheaper models automatically. Save 40-70% on AI API costs. </p> <p align="center"> <a href="https://pypi.org/project/nadirclaw/"><img src="https://img.shields.io/pypi/v/nadirclaw" alt="PyPI" /></a> <a href="https://github.com/doramirdor/NadirClaw/actions"><img src="https://github.com/doramirdor/NadirClaw/actions/workflows/ci.yml/badge.svg" alt="CI" /></a> <a href="https://pypi.org/project/nadirclaw/"><img src="https://img.shields.io/pypi/pyversions/nadirclaw" alt="Python" /></a> <a href="LICENSE"><img src="https://img.shields.io/github/license/doramirdor/NadirClaw" alt="License" /></a> <a href="https://github.com/doramirdor/NadirClaw"><img src="https://img.shields.io/github/stars/doramirdor/NadirClaw?style=social" alt="GitHub stars" /></a> </p> <p align="center"> Works with <strong>Claude Code</strong> · <strong>Cursor</strong> · <strong>Continue</strong> · <strong>Aider</strong> · <strong>Windsurf</strong> · <strong>Codex</strong> · <strong>OpenClaw</strong> · <strong>Open WebUI</strong> · Any OpenAI-compatible client </p> <p align="center"> <a href="https://getnadir.com?ref=readme-nav"><strong>Nadir Pro (hosted)</strong></a> · <a href="#quick-start">Quick Start</a> · <a href="#nadirclaw-vs-nadir-pro">OSS vs Pro</a> · <a href="docs/comparison.md">Comparisons</a> · <a href="https://github.com/doramirdor/nadirclaw-action">GitHub Action</a> </p> <p align="center"> <sub>Running real traffic or a team? <a href="https://getnadir.com?ref=readm

Topics

Explore more

Data from GitHub. Synced on 2026-06-17