eliasbiondo/antirion repository overview

⁠Antirion

The fastest open-source AI gateway. One Rust binary, with the admin console built in, between your apps and every model provider.

GitHub⁠ · Getting started⁠ · Benchmarks⁠ · Releases⁠

⁠Quick start

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

No config file is needed. The first start prints the console's address and an admin token, shown that once:

  Console      http://localhost:8081/ui/
  Admin token  ak-antirion-…
               Shown on this first start only; it is kept in
               /var/lib/antirion/data/bootstrap-token
  Data plane   http://localhost:8080/v1

1. Open http://localhost:8081/ui/⁠ and sign in with the admin token.
2. Add a provider (paste its API key), a model and an API key for your app.
3. Point any OpenAI- or Anthropic-compatible SDK at http://localhost:8080/v1 with that key.

curl http://localhost:8080/v1/chat/completions \
  -H "Authorization: Bearer $ANTIRION_KEY" -H "Content-Type: application/json" \
  -d '{"model": "<your model>", "messages": [{"role": "user", "content": "Hello"}]}'

Lost the token? It is kept in the volume: docker cp <container>:/var/lib/antirion/data/bootstrap-token - | tar -xO.

⁠What it does

• Fast. Sub-millisecond overhead at thousands of requests per second on 4 vCPUs, with an idle footprint of about 40 MiB. Rust, no garbage collector, one data thread per CPU (benchmarks⁠).
• Every major provider behind one API. 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 such as vLLM. Clients speak OpenAI, Anthropic, Gemini or Bedrock; Antirion translates between them.
• Routing that keeps you up. Fallback chains, weighted splits, least cost, least latency, retries, hedging and circuit breakers, decided per request and recorded in the request log.
• Spend under control. Rate limits and budgets in dollars or tokens per organization, team, project and key, with a cost ledger and alerts before a budget runs out.
• Safety at the edge. Guardrails for personal data, secrets, keywords, regexes, JSON schemas, topics, moderation and prompt injection, on requests, responses and streams.
• Ready for production. Exact and semantic caching, MCP and agent gateways, SSO with OIDC, SAML and SCIM, an audit log with change requests, and a standalone node or a cluster on PostgreSQL and Valkey.

⁠Tags

Images are multi-arch (linux/amd64, linux/arm64), distroless, run as UID 65532 with no shell, and are also published as ghcr.io/eliasbiondo/antirion:vX.Y.Z. Each release image is signed with Sigstore (keyless, bound to the release workflow) and carries an SBOM and SLSA provenance:

cosign verify eliasbiondo/antirion:0.1.1 \
  --certificate-identity-regexp '^https://github.com/eliasbiondo/antirion/.github/workflows/' \
  --certificate-oidc-issuer https://token.actions.githubusercontent.com

⁠Configuration

Data lives in the volume /var/lib/antirion; name it (as above) so a new container keeps the same gateway. docker run --rm eliasbiondo/antirion serve --help lists every flag.

To run from a config file, mount it and pass serve --config:

docker run -d --name antirion --read-only -p 8080:8080 -p 8081:8081 \
  -e ANTIRION_BOOTSTRAP_TOKEN -e ANTIRION_ENCRYPTION_KEY \
  -v "$PWD/antirion.yaml:/etc/antirion/antirion.yaml:ro" -v antirion-data:/var/lib/antirion \
  eliasbiondo/antirion serve --config /etc/antirion/antirion.yaml

The configuration reference, Docker Compose (with a two-node cluster on PostgreSQL and Valkey) and the Helm chart for Kubernetes are in the repository⁠.

⁠License

Apache-2.0. Source, issues and contributions: https://github.com/eliasbiondo/antirion⁠.