diegosouzapw/OmniRoute

Never stop coding. Free MIT AI gateway: one endpoint, 352 providers (150+ free), 1200+ models Kimi, Claude, GPT, Gemini, GLM, DeepSeek, MiniMax. Works with Claude Code, Codex, Cursor, OpenCode, Cline & Copilot. Quota-aware auto-fallback, RTK+Caveman compression saves 15-95% tokens, MCP/A2A, Desktop/PWA. Built by 550+ contributors

📦 What is OmniRoute?

OmniRoute is an open‑source AI gateway that lets you call any LLM provider through a single OpenAI‑compatible endpoint (e.g. http://localhost:20128/v1). It automatically selects the best backend – from paid subscriptions down to completely free tiers – and can fall back to another provider if the first one fails. The goal is to give developers a zero‑config, cost‑free way to use LLMs while keeping track of usage, quotas, and token‑saving compression.

🔑 Core Features (as described in the README)

Feature What it does
352 AI providers / 150+ free tiers OmniRoute maintains a catalog of 444 free‑tier entries across 34 recurring pool keys, exposing them through one endpoint.
Auto‑fallback routing 19 routing strategies (subscription → API key → cheap → free) keep requests alive even when a provider is down or rate‑limited.
Token compression (RTK + Caveman) Saves 15‑95 % of tokens (average ~89 %) before sending prompts, dramatically extending free‑tier budgets.
Live quota telemetry Dashboard (/dashboard/free-tiers) shows the current free‑token budget (~1.47 B tokens/month) and per‑provider usage.
Zero‑config start After npm i -g omniroute the server starts on localhost:20128; calling the auto model works out‑of‑the‑box with no API keys.
Multi‑modality (vision/audio/video) Added in v3.8.50 – the router can forward image, audio, or video payloads to providers that support them.
Private / local‑first mode Keys can be stored encrypted with AES‑256‑GCM, allowing teams to share a gateway without exposing secrets to the cloud.
CLI & MCP integration Comes with a full command‑line interface and a “Multi‑Client Proxy” that integrates with 36 existing CLIs/agents (e.g., Claude Code, Cursor, Cline).
Docker & Electron builds Available as an npm package, Docker image, and desktop app for easy deployment on any platform.
Extensive documentation & community Docs, quick‑start examples for Python/Node/PHP/cURL, a Discord/Telegram community, and a public roadmap.

🚀 How it works (high‑level flow)

  1. Install – npm i -g omniroute (or pull the Docker image). The server starts locally.
  2. Point your tool – configure any OpenAI‑compatible client to http://localhost:20128/v1.
  3. Send a request – use the special model name auto (or oc/... for a specific free backend).
  4. Router decides – OmniRoute checks quota, health, and cost, applies compression, then forwards the request to the chosen provider.
  5. Response returns – the client receives a standard OpenAI‑style JSON reply, unaware of the underlying provider.

📊 Who might benefit?

  • Developers & hobbyists who want to experiment with many LLMs without managing dozens of API keys.
  • Teams looking to share a single gateway while keeping usage transparent and costs at $0.
  • Tool makers (IDE extensions, CLI agents) that need a stable endpoint regardless of provider outages.
  • Researchers interested in benchmarking many models under a uniform API.

📚 Where to learn more

  • Quick‑start scripts: examples/quickstart/ (Python, Node, PHP, cURL).
  • Free‑tier methodology: docs/reference/FREE_TIERS.md.
  • Roadmap & changelog: ROADMAP.md and the version table in the README.
  • Community: Discord, Telegram, WhatsApp groups, and the public website omniroute.online.

🛠️ Tech stack (as listed)

  • Node.js runtime (npm package omniroute).
  • Docker container for server deployment.
  • Electron for a desktop UI.
  • RTK + Caveman compression algorithms.
  • AES‑256‑GCM encryption for local‑first key storage.

📦 Install options (summary)

  • npm: npm i -g omniroute → runs on localhost:20128.
  • Docker: docker pull diegosouzapw/omniroute.
  • Source: clone the repo and run via pnpm/npm.
  • Electron: download from GitHub releases.

🎉 Bottom line

OmniRoute is a real, production‑grade gateway that aggregates hundreds of LLM providers, automatically optimizes for cost and availability, and offers a plug‑and‑play OpenAI‑compatible API. It is squarely in the AI tooling space and provides tangible value for anyone wanting to use LLMs without the hassle of managing multiple keys or paying for usage.

Written about in

Related

  • Project
    theopenco/llmgateway

    LLM Gateway is an open‑source API gateway that provides a single OpenAI‑compatible endpoint for routing requests to multiple LLM providers, centralising API‑key management and offering built‑in usage and performance analytics. It can be used as a hosted SaaS service or self‑hosted via Docker.

  • Project
    tashfeenahmed/freellmapi

    A local router that aggregates free tiers from dozens of LLM providers into a single OpenAI-compatible endpoint, providing billions of tokens of free inference capacity.

  • Project
    Continuum-AI-Corp/OrcaRouter-Lite

    A self-hosted, OpenAI-compatible LLM router that optimizes costs and reliability using automatic model selection and a managed fallback safety net.

  • Project
    CyberStrikeus/CyberStrike

    CyberStrike is an open‑source AI agent that turns any LLM (Claude, GPT, Gemini, etc.) into an autonomous penetration‑testing platform. It adds a security‑focused “intelligence layer”, ships 13+ specialised agents and 7,600+ signed security‑skill files, supports 150+ AI providers (including offline models), and can orchestrate real tools locally or on remote Bolt servers. A TUI, optional web UI, and built‑in HackBrowser complete the workflow.

  • Project
    ulab-uiuc/LLMRouter

    An open-source library for intelligent LLM routing that dynamically selects the most suitable model for each query to optimize inference cost and performance.