One endpoint for selected AI models.

Quick start

Create a key, choose a model, send a request

API keys are created in the user dashboard and shown in full only once.

• Create one project and API key per application or environment.
• Choose an Available model from the live catalog.
• Paid models use bare IDs; free models use the free/ prefix.

curl https://api.eidosstack.com/v1/chat/completions \
  -H "Authorization: Bearer $EIDOSSTACK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"deepseek-v4-flash","messages":[{"role":"user","content":"Hello"}],"max_tokens":256,"stream":true}'

Open API keys · View models and pricing

API

Available endpoints

Streaming

Receive tokens as they arrive

Set stream: true. The response uses Server-Sent Events and ends with [DONE].

• Set max_tokens to cap output length.
• Do not automatically retry a stream after tokens have been emitted.
• For retryable non-stream requests, send an Idempotency-Key.

Error handling

Separate request, balance, limit, and upstream errors