About this project

free-llm-api-stack is a lightweight Python router designed to aggregate multiple free LLM API tiers—such as Groq, Cerebras, Google Gemini, and OpenRouter—into a single, resilient endpoint. It provides an OpenAI-compatible interface that features automatic failover; if a provider returns a rate limit error (HTTP 429), a server error (5xx), or a timeout, the request is transparently passed to the next available provider in the stack. The tool is built with zero external dependencies and uses a centralized providers.json file to manage API endpoints, environment variable names, and model limits. This architecture allows developers to maintain high availability for LLM-powered applications without exceeding individual free-tier quotas. Additionally, the repository includes a specialized skill for Claude Code to help automate the integration of resilient free backends into existing projects.