Gateways & routing
LLM gateways, proxies and model routers
diegosouzapw/OmniRoute OmniRoute is the clear category leader; the most important thing to know is it exposes a single local endpoint that auto-fallbacks across hundreds of providers with quota‑aware routing and stacked token compression, so it’s the easiest way to avoid provider limits and interruptions. If you prefer a much smaller operational footprint, an OpenAI-compatible drop-in, or specialized integrations (IDE agents, Qwen automation, or human-in-the-loop), several focused projects trade breadth for simplicity or niche features.
Which one matches your setup?
Answer any of the questions — the shortlist updates as you go. Recommendations come from the capability passports below, nothing else.
Comparison matrix
Axes are extracted from each project's docs by our review pipeline; the maturity score is computed from stars, growth and commit activity — not an opinion. Click a column to sort.
| Runs in | Models | Context | Cost to run | |||||
|---|---|---|---|---|---|---|---|---|
⭐ 78.7k +2794/7d | CLIIDECoding-agent pluginWeb app | BYOKOpenAIAnthropicGeminiFixed provider | Endpoint only | Guardrails & evals | ●●●●● | Self-hostable | ●●●●● | Free to start; includes keyless free providers |
⭐ 61.6k +508/7d | CLIWeb app | BYOK | No repo context | Basic guardrails | ●●●●● | Self-hostable | ●●●●● | Self-hosted — your API keys (pay providers) |
⭐ 13.2k +55/7d | CLIWeb app | BYOKOpenAIAnthropicGeminiFixed provider | Diff only | Guardrails & filters | ●●●●● | Self-hostable | ●●●●● | Uses your provider API keys; self-hosted or hosted enterprise plans. |
⭐ 624 | CLIWeb appCoding-agent pluginIDE | BYOKOpenAIAnthropicOllama / localFixed provider | Diff only | Basic filters | ●●●●● | Runs fully local | ●●●●● | Your API keys for upstream providers; local models run locally (no provider fees). |
⭐ 750 +1/7d | CLI | Fixed provider | Diff only | Routing + tags | ●●●●● | Self-hostable | ●●●●● | Free, self-hosted (no paid cloud required) |
⭐ 279 | CLIWeb app | BYOK | Message-only | Competence gating | ●●●●● | Self-hostable | ●●●●● | Your API keys/subscriptions or local models |
⭐ 745 | CLIIDECoding-agent plugin | Fixed provider | Diff/file only | No noise controls | ●●●●● | Third-party cloud | ●●●●● | Requires a Duel API key (dashboard subscription) |
⭐ 204 +1/7d | CLIWeb app | Fixed provider | Diff only | Content filter + guards | ●●●●● | Self-hostable | ●●●●● | Free via Qwen accounts; self-hosted |
⭐ 204 +1/7d | CLIWeb app | Fixed provider | Diff only | Content filter + spam guard | ●●●●● | Self-hostable (Qwen) | ●●●●● | Free via your Qwen accounts (self-hosted) |
⭐ 46 | CLIWeb app | BYOKOpenAIAnthropicGemini | No repo access | Confidence gating & audit | ●●●●● | Self-hosted gateway | ●●●●● | Your providers' API keys; provider billing applies |
⭐ 28 | Web appCLI | BYOK | Request context only | Governance gating | ●●●●● | Fully self-hostable | ●●●●● | Free (open-source, self-hosted) |
Capability profiles
Six axes, 0–5 each. The shape tells you the strategy: a wide hexagon is a generalist, a spike is a specialist. Showing the 8 most established — the rest are in the full catalog.
Provides a single local endpoint that auto-fallbacks across 290+ providers with quota-aware routing and stacked token compression (RTK+Caveman) so you rarely hit limits while saving tokens.
A lightweight, production-ready self-hosted AI gateway that unifies 100+ LLM providers into a single OpenAI-compatible API with virtual keys, spend tracking, guardrails, and load balancing.
A tiny, high-performance LLM gateway that routes to 1,600+ models while providing built-in guardrails, retries, and load‑balancing.
Multi-provider gateway that routes Claude Code and other coding agents across a broad provider catalog with ordered fallbacks and local-model runtimes.
Lets you run Hermes Agent, OpenClaw, and OpenCode simultaneously on a single WeChat account by acting as the sole iLink poller and proxying requests to local gateway endpoints.
A local model router that stages long-horizon agent work and uses competence scoring and bounded failover to route routine turns to cheaper models while protecting quality.
An IDE-native routing layer that runs prompts across multiple models and picks the cheapest answer that still wins.
A drop-in OpenAI-compatible gateway that uses browser-automated Qwen accounts to provide free Qwen models with multi-account rotation, session pooling, and streaming support.
All repositories (11)
Never stop coding. Free MIT AI gateway: one endpoint, 375 providers (150+ free), 1,400+ models Kimi, Claude, G…
The fastest, litest AI Gateway. Rust core with Python SDK. Call 100+ LLM APIs in OpenAI (or native) format wit…
A blazing fast AI Gateway with integrated guardrails. Route to 1,600+ LLMs, 50+ AI Guardrails with 1 fast & fr…
Open-source multi-provider AI gateway for Claude Code and other coding agents, with model routing, streaming,…
Make Codex and Claude Code subscriptions go further: a local model router that gives routine agent work to eco…
Drop-in OpenAI-compatible API gateway for Qwen AI models. Use your Qwen account (chat.qwen.ai) as a free AI AP…
Drop-in OpenAI-compatible API gateway for Qwen AI models. Use your Qwen account (chat.qwen.ai) as a free AI AP…
Enchant Version of 9Router. 304+ Providers (API-key, OAuth, free-tier, and 39 web-cookie providers), 6 combo s…
Human as Agent（人即智能体）· Human-as-LLM 人工代理网关：把工程师变成模型，OpenAI 兼容 /v1 接入 Agent 调度池，涉密/需人工任务路由给真实工程师。为 AI Agent 提…
Hosted alternatives
If running your own reviewer is more ops than you want, these managed services cover the same job.
A hosted gateway to hundreds of models behind one API, with routing and fallbacks handled, so you swap providers without running your own proxy.
Try OpenRouter →Managed AI gateway with routing, caching, retries and observability, the hosted counterpart to running an open-source gateway in your stack.
Try Portkey →A hosted unified endpoint across providers with fallbacks and usage tracking, so multi-model access is managed rather than proxied by you.
Try Vercel AI Gateway →