What is this?
OmniRoute is a free, open-source AI gateway that puts 359 providers and 1200+ models (Kimi, Claude, GPT, Gemini, GLM, DeepSeek, MiniMax, etc.) behind one API endpoint, with quota-aware automatic fallback when a provider runs out.
OmniRoute is a free, open-source AI gateway that puts 359 providers and 1200+ models (Kimi, Claude, GPT, Gemini, GLM, DeepSeek, MiniMax, etc.) behind one API endpoint, with quota-aware automatic fallback when a provider runs out.Is this a hosted service or something I run myself?You self-host it (npm or Docker Compose) and it exposes an OpenAI/Anthropic-compatible endpoint plus a desktop/PWA client; your coding tools point at that endpoint instead of calling each provider's API directly.
You self-host it (npm or Docker Compose) and it exposes an OpenAI/Anthropic-compatible endpoint plus a desktop/PWA client; your coding tools point at that endpoint instead of calling each provider's API directly.How do I get it running?Install with npm (which generates a local .env during postinstall) or run it via Docker Compose, then point your CLI or app at the gateway endpoint — it works with zero config right after install using built-in free backends.
Install with npm (which generates a local .env during postinstall) or run it via Docker Compose, then point your CLI or app at the gateway endpoint — it works with zero config right after install using built-in free backends.Do I need my own OpenAI or Claude API key to use it?No — it works the moment you install it with no keys or config, calling a free OpenCode backend by default; you can add your own provider keys later for paid models.
No — it works the moment you install it with no keys or config, calling a free OpenCode backend by default; you can add your own provider keys later for paid models.Is the ~1.62B free tokens/month figure guaranteed?No — that headline is re-audited every two weeks against the live provider catalog and can drop if a free tier ends; quotas requiring regional identity checks are shown separately and never counted into it.
No — that headline is re-audited every two weeks against the live provider catalog and can drop if a free tier ends; quotas requiring regional identity checks are shown separately and never counted into it.Which coding tools and CLIs does it plug into?It integrates directly with Claude Code, Codex, Cursor, OpenCode, Cline, Copilot, Aider, Goose and Qwen Code, with one-command setup via its own `omniroute run` and `omniroute configure` commands.
It integrates directly with Claude Code, Codex, Cursor, OpenCode, Cline, Copilot, Aider, Goose and Qwen Code, with one-command setup via its own `omniroute run` and `omniroute configure` commands.What is a 'combo' in OmniRoute?A combo is a chain of models OmniRoute routes across automatically — if one provider runs out of quota, fails, or spikes in cost, requests move to the next eligible model in the chain.
A combo is a chain of models OmniRoute routes across automatically — if one provider runs out of quota, fails, or spikes in cost, requests move to the next eligible model in the chain.What does the RTK+Caveman compression do?It's a built-in prompt/context compression layer that OmniRoute claims cuts token usage by 15-95%, with language-specific packs (DE/FR/JA/Chinese) and filters for things like Gradle and .NET output.
It's a built-in prompt/context compression layer that OmniRoute claims cuts token usage by 15-95%, with language-specific packs (DE/FR/JA/Chinese) and filters for things like Gradle and .NET output.What is OmniRoute Radar?Radar is an optional, opt-in, GET-only catalog overlay that gives supporters fresher free-model availability data; OmniRoute states its client never uploads prompts, traffic, or usage telemetry.
Radar is an optional, opt-in, GET-only catalog overlay that gives supporters fresher free-model availability data; OmniRoute states its client never uploads prompts, traffic, or usage telemetry.What are similar projects?9router, casbin-gateway, ThinkWatch-Lite, busbar, opencodex, gpt-load, openclaw-zero-token, magpie
9router, casbin-gateway, ThinkWatch-Lite, busbar, opencodex, gpt-load, openclaw-zero-token, magpieWhich topics is it in?free-ai, cline, multi-provider, llm-proxy, a2a, llm-gateway, ai-gateway, kimi, load-balancer, compression, api-gateway, copilot
free-ai, cline, multi-provider, llm-proxy, a2a, llm-gateway, ai-gateway, kimi, load-balancer, compression, api-gateway, copilot