Skip to content

One API for the models you build on

One key.
Every model.

Run your AI workloads through a single API, and let Space Router handle the providers, the failover and the spend limits.

  • OpenAI-compatible
  • Pay as you go
  • A budget on every key

Quality, speed, cost.
Stop choosing one.

Send a request without naming a model and Space Router picks one for the job, so easy questions go to cheap models and hard ones go to capable ones.

A provider’s bad day
stays their problem.

When a provider times out or hits a rate limit, the request moves to another one serving the same model. You write no retry logic.

Every request, model
and dollar on one screen.

Usage and spend are recorded against the key that made the request. Give each team a budget and see who is close to theirs.

Your rules, on every request

Decide which models and providers each team may use and how much they may spend. The limits apply whatever model is called.

  • Budgets per team and key
  • Rate limits per key
  • Model allow-lists
  • Provider allow-lists
  • Bring your own keys
  • Usage logs
  • Single sign-on
  • Key rotation

Start with one request.

Your first request takes a few minutes. Pay for what you use, with no subscription.

curl https://api.spacerouter.ai/v1/chat/completions \
  -H "Authorization: Bearer $SPACEROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic/claude-sonnet-4.6",
    "messages": [
      {"role": "user", "content": "Hello from orbit"}
    ]
  }'