Generate

Send a prompt to Aivana MMI and get one synthesized answer back.

POST /v1/generate is the main endpoint of the Aivana API. You send a prompt; Aivana decides how much intelligence the request needs, uses one or several AI models as appropriate, reconciles their work, and returns a single answer. You don't choose or manage the models.

POST https://api.aivana.ai/v1/generate

Send your API key in the X-API-Key header on every request. Create a key on the API Keys page. The secret is shown only once; keep it on your server, in a secret manager or an environment variable, and never in browser or mobile code.

Only a prompt is needed. Every other option is optional; see Parameters.

curl https://api.aivana.ai/v1/generate \
  -H "X-API-Key: $AIVANA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Write a haiku about the ocean."}'

For a single question, send prompt (up to 200,000 characters). For a multi-turn exchange, send messages instead: up to 40 items, each with a role of system, user or assistant and a content string, oldest first. See Conversations.

{
  "messages": [
    { "role": "system", "content": "You are a concise support lead." },
    { "role": "user", "content": "A customer wants a refund after 40 days." }
  ]
}

A successful call returns 200 with the answer and what the request used. This example is abridged; Responses & Usage describes every field.

{
  "id": "gen_...",
  "request_id": "req_...",
  "answer": "Waves fold and unfold, ...",
  "usage": { "input_tokens": 14, "output_tokens": 22, "cached_tokens": 0, "credits": 0.0001 },
  "latency_ms": 2140,
  "finish_reason": "stop",
  "notices": []
}

curl https://api.aivana.ai/v1/generate \
  -H "X-API-Key: $AIVANA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "What is the default port for Postgres?", "effort": "low"}'

The JSON Schema (Strict) request returns the validated object in structured, or an error: see Structured Outputs.