antirion

πŸ•·οΈ The fastest open-source AI Gateway and LLM Gateway β€” one binary with built-in console to route, secure, budget and observe every LLM call across OpenAI, Anth

What is antirion?

πŸ•·οΈ The fastest open-source AI Gateway and LLM Gateway β€” one binary with built-in console to route, secure, budget and observe every LLM call across OpenAI, Anthropic, Gemini, Bedrock and more.

Overview of antirion

What it is

Antirion is an open-source AI gateway and LLM gateway written in Rust. It is distributed as a single binary that includes a built-in admin console. The gateway sits between applications and model providers, routing, securing, budgeting, and observing every LLM call. It supports providers such as OpenAI, Anthropic, Gemini, Bedrock, and many others. According to its README, it is the fastest open-source AI gateway measured.

Key features

  • Performance: Sub-millisecond overhead at thousands of requests per second on 4 vCPUs. In benchmarks, it achieved a p50 overhead of 0.53 ms and a throughput of 21,512 req/s over 64 concurrent connections on a c7i.xlarge machine.
  • Single binary with console: The gateway, Admin API, and web console ship as one Rust binary (or a distroless image). No sidecars, separate UI deployment, or runtime installation required.
  • Unified API for major providers: Supports OpenAI, Azure OpenAI, Anthropic, Amazon Bedrock, Google Vertex AI, Gemini, Mistral, Cohere, Groq, Together, Fireworks, DeepSeek, xAI, Cerebras, DeepInfra, Perplexity, and any OpenAI- or Anthropic-compatible server (e.g., vLLM). Clients can speak OpenAI, Anthropic, Gemini, or Bedrock protocols; Antirion translates between them.
  • Advanced routing: Fallback chains, weighted splits, least cost, least latency, least busy, tiers, content classifiers, retries, hedging, and circuit breakers. Routing decisions are made per request and recorded in the request log.
  • Spend control: Rate limits and budgets in dollars or tokens at the organization, team, project, and key levels. Prices per model, a cost ledger, and alerts before a budget runs out.
  • Safety at the edge: Guardrails for PII, secrets, keywords, regexes, JSON schemas, tool permissions, topics, moderation, model-based judges, and partner services. Supports request, response, and stream inspection with redaction that preserves format.

Use cases

  • Unified API gateway: Replace multiple provider-specific SDKs with a single endpoint that handles routing, fallback, and translation between protocols.
  • Production monitoring and control: Observe every LLM call through the built-in request log, enforce rate limits and budgets, and set up alerts.
  • Team collaboration: Invite team members via the console, use SSO with OIDC, SAML, and SCIM, and manage API keys with granular permissions.
  • High-performance scenarios: Suitable for applications requiring low latency and high throughput, such as real-time chat or large-scale inference pipelines.

Getting started

docker run -p 8080:8080 -p 8081:8081 -v antirion-data:/var/lib/antirion eliasbiondo/antirion

License and pricing

Antirion is open-source software licensed under Apache-2.0. It is free to use and self-host.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/eliasbiondoantirion.svg)](https://yunodb.com/tools/eliasbiondoantirion?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/eliasbiondoantirion?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/eliasbiondoantirion.svg" alt="Featured on YunoDB" height="20"></a>