Apache-2.0 · self-host the binary

Agent-native LLM router that optimizes your agent with every run.

Zero harness changes — every model call reliable, traceable, secure, and cost-effective. Open-sourced, Cloud opt-in.

$curl -fsSL https://bitrouter.ai/install.sh | sh
claude-code — bitrouter
opus·sonnet·gemini

Where AI agents break.

Coding agents run into these first. Every autonomous agent runs into them eventually.

A blip at file 140 shouldn’t kill the run.

Your coding agent is 140 files into a refactor when a provider rate-limits it, and the run dies. Every long agent loop breaks the same way — one blip, back to step zero.

  • Overnight jobs that finish
  • No babysitting rate limits
  • The agent never sees the failure
agent · run_8x2k

Why agents run on BitRouter.

Four mechanisms, built into the router — not bolted on per agent.

01 Reliability

One provider fails. Your agent run doesn’t.

Reroutes across providers mid-run, transparently — your agent never sees the failed call.

Powered byIntent-aware routingMulti-provider failover
router · run_8x2k
02 Observability

Trace every hop. Not just every request.

Full call-chain visibility: every agent, every model, every step, with cost attributed per run.

Powered byPer-run cost attributionFull call-chain traces
trace · run_8x2k9hf3
03 Security

Guardrails for every agent. Configured once.

Prompt-injection detection, output filtering, and rate limits — enforced at the router, once, for every agent.

Powered byRouter-level guardrailsPer-agent identity
policy · router
04 Efficiency

Not every call needs your strongest model.

Matches each call to the right model by task complexity — savings that compound across every run.

Powered byModel-per-task routingPrice-aware model selection
router · cost model

Questions before you ship.

Common questions about pricing, routing, and data handling. If yours isn’t here, check the docs or talk to us — we usually reply within a day.

OpenRouter is a closed-source hosted gateway — no self-host option, no agent-native primitives, no permissionless registry. BitRouter is Apache 2.0: fork the binary and run it anywhere, or use the hosted edge if you don’t want to operate it. The provider registry is open — anyone can publish a provider via PR, no review queue.
LiteLLM is an open-source Python library you embed in your app. BitRouter is a standalone binary with auth, billing, observability, guardrails, and an MCP/ACP/Skills gateway built in. Drop it in front of any runtime — Claude Code, Cursor, your own agent — without wiring it into every service.
Every routing alias retries with exponential backoff and fails over cross-provider on 429s or upstream errors — a single provider outage doesn’t kill an agent run. Per-provider health and incident history are visible in the console. Failed requests aren’t billed.
Added on top of the upstream provider’s published rate — 2% with USDC/USDT over x402/MPP, 5% with card (Stripe’s 3% fee absorbed). No token markup, no minimums, no monthly fee. Self-host has no platform fee at all.
Open a PR against bitrouter/provider-registry with a YAML listing your name, base URL, the models you serve, prices, and routing weight. CI runs schema checks; once green, BitRouter picks up your listing on the next refresh and starts routing agent traffic. No application, no gatekeeper.
Pull the Apache 2.0 binary from github.com/bitrouter/bitrouter — single binary, no daemon, no GUI, no infrastructure dependencies. Drops into any container or CI step. You get the same routing, guardrails, MCP/ACP/Skills gateway, and observability as the hosted edge — without the platform fee.

Start routing in under a minute.

One API key, every model. No monthly fees, no lock-in. Self-host or use the hosted edge.

$curl -fsSL https://bitrouter.ai/install.sh | sh
All systems operational·2% on usage · no platform fee < $10/mo·Apache-2.0