Tag
Gateway tools
8 reviewed tools: 7 free or open source, 1 hosted with a real free tier.
Order: free and open source first, then hosted products with a real free tier, then paid-led; catalog order inside each group, with quiet or archived repos last. Not a score, and no tool pays for its place.
8 tools
- Janus GatewayGPL-3.0 plugin-core WebRTC gateway (SIP, VideoRoom, streaming)
Free / OSS · GPL-3.0 · 9.2k GitHub stars
VerdictGPL-3.0 WebRTC gateway. Self-host free; commercial options exist — pitch the OSS core.
CatchPlugin-core C gateway — not LiveKit/mediasoup opinionated rooms products. You pick plugins and ops the gateway. Distinct from LiveKit, mediasoup, PeerJS, and Pion.
- BifrostApache-2.0 AI gateway — perf angle vs LiteLLM (Maxim)
Free / OSS · Freemium · 9k GitHub stars
VerdictApache-2.0 AI gateway. Fast LiteLLM alternative; Maxim enterprise is the loud paid path.
CatchVs LiteLLM: same gateway job with a Go/perf positioning — not a model marketplace. Maxim upsell: enterprise/clustering/adaptive LB/OIDC provisioning and getmaxim.ai Bifrost enterprise pages are the commercial gravity; OSS gateway still runs locally. You still pay upstream model providers.
- ContextForgeApache MCP+A2A+REST/gRPC→MCP gateway with governance + OTel
Free / OSS · Apache-2.0 · 4.7k GitHub stars
VerdictApache MCP/A2A gateway when you need federation + governance + OTel on your own infra and will pay the ops cost; skip if you only needed a thin remote MCP shim.
CatchVs SuperGateway and mcp-remote: ContextForge is a heavier self-host governance/observability gateway, not a thin MCP bridge. Vs Smithery: this is a self-host gate you operate, not a hosted MCP marketplace. Open-issue count is high for the maturity pitch; local/JWT setups need real secrets — do not ship example tokens.
- MetaMCPMIT self-host MCP aggregator + middleware + inspector (one endpoint)
Free / OSS · OSS/MIT · 2.7k GitHub stars
VerdictMIT self-host MCP aggregator when one namespaced endpoint beats wiring every server — only with eyes open on thin maintainer capacity and the `ai-dev` default branch.
CatchVs MCP Registry: listing/discovery, not an aggregator runtime. Vs ContextForge: protocol gateway + A2A/OTel plane. Vs Klavis: freemium integrations/Strata. Vs Smithery: hosted marketplace. Vs SuperGateway: transport stdio↔SSE only. Soft maintainer capacity thin — last push 2026-06-22; default branch is `ai-dev` (caution); author says core still works / will keep merging PRs — shippable with eyes open.
- Claude Code RouterMIT local control plane — route coding agents across models
Free / OSS · OSS/MIT · 37.8k GitHub stars
VerdictMIT local agent model router when you want CCR’s gateway/UI — not Weave ELv2, not HarnessRouter; install `@musistudio/claude-code-router` only.
Catch★37492 in about 83 weeks (~451 a week): judge it on the router/gateway loop, not the star count. Vs weave-router / harnessrouter: Weave is ELv2 on-box embedder/prompt routing; HarnessRouter CE is an Apache unified API over coding harnesses — CCR’s catch is a local Claude-Code-oriented model router gateway + UI. Pin the scoped `@musistudio/` package name; bare npm name is a different project.
- HigressApache-2.0 AI and API gateway on Envoy and Istio, with Wasm plugins, model proxying and MCP hosting
Free / OSS · Apache-2.0 · 9.6k GitHub stars
VerdictA production-grade AI gateway if you can run Envoy and Istio — for a quick model proxy, start with something smaller.
CatchIt is a full Envoy/Istio gateway, much heavier to run than a single-binary model proxy. Default images pull from an Alibaba Cloud registry in Hangzhou, so outside China switch to the US or Southeast Asia mirror if pulls time out. Much of the documentation and the plugin catalog live on higress.cn. Vs agentgateway (a lighter, purpose-built gateway for agent and MCP traffic) and litellm (a Python model proxy you can start in minutes): Higress fits when you want AI traffic inside a production API gateway you already run at scale.
- OtariApache-2.0 self-hosted OpenAI-compatible LLM gateway
Free / OSS · Apache-2.0 · 711 GitHub stars
VerdictA clean, self-hosted way to put keys and budgets in front of many LLM providers — pin a version while it is pre-1.0.
CatchIt is younger than LiteLLM and still pre-1.0, so expect breaking changes. The one-line Docker quick start keeps SQLite inside the container and loses everything when it stops; a persistent setup means Compose with PostgreSQL. The README also offers a hosted control plane at otari.ai, so check which features you need run fully standalone.
- LiteLLMMIT OpenAI-compatible gateway to 100+ LLMs; Enterprise SSO paid
Free tier · Freemium · 61.7k GitHub stars
VerdictMIT LLM gateway. Self-host free; Enterprise SSO/governance is paid.
CatchVs OpenRouter/provider SDKs: self-host gateway you operate — not a hosted model marketplace. Enterprise/ tree and SSO/audit/RBAC are commercial; core proxy stays MIT. You still pay model providers.
FAQ
- What is the top free pick among Gateway tools?
- Janus Gateway leads the Stack Gems list for Gateway tools. Verdict: GPL-3.0 WebRTC gateway. Self-host free; commercial options exist — pitch the OSS core.
- How many Gateway tools does Stack Gems review?
- 8: 7 free or open source, 1 hosted with a real free tier. Each review has a verdict and an honest Catch; no tool pays for its spot.
- Which Gateway tools can I self-host for free?
- Janus Gateway, ContextForge, MetaMCP, Claude Code Router, Higress and Otari.
- Which Gateway tools have the most GitHub stars?
- LiteLLM (61.7k), Claude Code Router (37.8k), Higress (9.6k), per GitHub on 2026-10-21. Stars measure attention, not fit; read the Catch.
- What is the catch with Janus Gateway?
- Plugin-core C gateway — not LiveKit/mediasoup opinionated rooms products.