SSC-STUDIO / Ai-Model-Gateway

AI Model Gateway

Self-hosted LLM operations gateway for teams that want routing, fallback, telemetry, benchmarks, config publishing, diagnostics, updates, and rollback inside their own environment.

View on GitHub Evaluate in 15 minutes Docker Compose Adoption checklist Client integrations Codex CLI gateway Claude Code gateway Run fallback demo Share kit Chinese share kit 中文评估页

If this matches your self-hosted LLM infrastructure needs, star the repository so more operators can find it.

Built for operators

Route LLM traffic without handing over the control plane.

Provider fallback

Route around quota, timeout, and upstream failures while keeping provider keys and policy local.

Config publish

Preview, diff, publish, audit, and roll back routing changes instead of editing a live proxy file.

Telemetry and cost

Inspect traffic, latency, model usage, provider health, request logs, and cost signals in one place.

Benchmark before routing

Compare models with exact, judge, JSON, tool, and stream scoring before promoting traffic.

Executable proof

Fallback behavior you can test locally.

The provider fallback demo starts two fake OpenAI-compatible upstreams. The primary returns 429, the gateway serves the fallback provider, rewrites the forwarded model, and records route_mode=model_fallback.

go test ./examples/provider-fallback -run TestProviderFallbackDemo -v
Open the provider fallback demo
AI Model Gateway monitoring workspace showing operational metrics

Review evidence

Check installability, quality, and security before adopting it.

Release archive install

Try the packaged v1.4.4 runtime with checksum verification, local config, runtime directories, and supervised startup commands.

Open release install path

Quality evidence

Review CI gates, local reproduction commands, runtime smoke checks, feature proof points, and current capability boundaries.

Open quality evidence

Security and trust model

Inspect admin auth, same-origin browser writes, provider-key handling, SSRF defenses, telemetry sensitivity, and update trust.

Open security model

Evaluation path

Start with the shortest useful trial.

  1. Check the fit. Use cases, self-hosted checklist, and comparison guide.
  2. Start the runtime. Build the compact Go runtime and open the Admin UI.
  3. Verify fallback. Run the executable demo against fake upstreams.
  4. Inspect operations. Review config publish, rollback, provider health, and roadmap docs.
15-minute evaluation path Docker Compose trial LLM gateway adoption checklist Self-hosted LLM gateway page Client integrations page Codex CLI gateway page Claude Code gateway page OpenAI-compatible gateway page OpenAI Anthropic gateway page Provider fallback gateway page LLM gateway comparison page Chinese self-hosted page Share kit Chinese share kit LLM gateway comparison Project roadmap

Support discovery

Help the right operators find it.

If AI Model Gateway fits your self-hosted routing, fallback, telemetry, or config rollback workflow, star the repository after evaluation. Feedback is also useful if something blocks adoption.

Star on GitHub Leave operational feedback
AI Model Gateway overview workspace AI Model Gateway operations workspace on mobile AI Model Gateway benchmark workspace on mobile