{"id":"omniroute","name":"OmniRoute","url":"https://wunderlust.xyz/r/omniroute","category":"agents","category_label":"Agents & dev","tags":["free","open-source","mcp"],"short":"An open source gateway that routes AI tools through an OpenAI-compatible endpoint","what":"An AI gateway with a local OpenAI-compatible endpoint that directs calls to hundreds of providers and models.","uses":"Use it when you want to connect Claude Code, Codex, Cursor or Cline to several models, with automatic fallback when a limit runs out.","benefits":["Automatic quota-based fallback between providers, with circuit breakers and key cooldown","Combined routing strategies and auto IDs that pick the route","MCP server, a CLI to launch supported tools, and a dashboard; keys stored encrypted and locally"],"caveats":["MIT license; free access depends on third parties' free plans, with each provider's limits","According to the project, token compression saves 15% to 95%; this is a maker claim"],"links":{"official":"https://github.com/cheaper-inference/OmniRoute"},"image":"https://wunderlust.xyz/media/omniroute-poster.jpg","video":"https://wunderlust.xyz/media/omniroute.webm","share_image":"https://wunderlust.xyz/og/r/omniroute.jpg","similar":["obsidian-local-rest-api","agent-beacon","open-notebook","unsloth-studio","artemis","playwright"],"added_on":"2026-11-05","pt":{"name":"OmniRoute","short":"Gateway open source que roteia ferramentas de IA por um endpoint compatível com OpenAI","what":"Gateway de IA com um endpoint local compatível com OpenAI que direciona as chamadas para centenas de provedores e modelos.","uses":"Usar quando você quer ligar Claude Code, Codex, Cursor ou Cline a vários modelos, com fallback automático quando um limite estoura.","benefits":["Fallback automático por cota entre provedores, com circuit breakers e cooldown de chaves","Estratégias de roteamento combinadas e IDs auto que escolhem a rota","Servidor MCP, CLI para abrir ferramentas suportadas e dashboard; chaves guardadas criptografadas e localmente"],"caveats":["Licença MIT; o acesso gratuito depende dos planos gratuitos de terceiros, com limites de cada provedor","Segundo o projeto, a compressão de tokens economiza de 15% a 95%; é alegação do fabricante"]},"schema_version":1}