local-optimind-api

Local OptiMind API is an open-source, locally-run gateway that routes LLM requests across multiple AI providers (OpenAI, DeepSeek, Anthropic, Groq, Together, local LLMs) using a single OpenAI-compatible endpoint. It intelligently selects the optimal provider for each request based on configurable modes: cost-first, quality-first, or balanced. The system includes automatic model discovery, custom pricing tiers, health tracking with exponential backoff, request logging, and SSE streaming support. An admin dashboard allows managing providers, models, pricing, and viewing logs via web UI. Clients are available in Python, JavaScript/TypeScript, Go, Java, and .NET with identical APIs. The router supports fallback routing on failure, task-based model inference when exact matches aren't available, and integrates with Docker for easy deployment. Architecture includes health filtering, model matching, scoring logic, and execution layers. Configuration is managed via environment variables. MIT licensed, with contributions welcome for new provider adapters, token counting, and routing strategies.