Skip to main content

Generate Completion

Generate a completion response for a given prompt. This is a streaming endpoint by default.

POST /api/generate

Request Parameters​

  • model (string, required): Model tag to execute (e.g. gemma4).
  • prompt (string, required): Input prompt text.
  • suffix (string, optional): Text appended after completion.
  • images (array of strings, optional): Base64-encoded image files.
  • stream (boolean, default true): Return a stream of JSON chunks.
  • raw (boolean, optional): Bypass template formatting logic.
  • format (string/object, optional): Set to "json" or provide a JSON Schema structure.
  • options (object, optional): Override Modelfile parameters (e.g. {"temperature": 0.5}).
  • keep_alive (string/number, default "5m"): Cache duration in memory.

Examples​

1. Streaming Request​

curl http://localhost:11434/api/generate -d '{
  "model": "gemma4",
  "prompt": "Why is the sky blue?"
}'

Chunk Payload​

{
  "model": "gemma4",
  "created_at": "2026-06-05T14:46:00Z",
  "response": "The",
  "done": false
}

Final Chunk Payload​

{
  "model": "gemma4",
  "created_at": "2026-06-05T14:46:05Z",
  "response": "",
  "done": true,
  "total_duration": 4820918000,
  "load_duration": 12093000,
  "prompt_eval_count": 12,
  "prompt_eval_duration": 18239000,
  "eval_count": 210,
  "eval_duration": 4790290000
}

2. Non-Streaming Request​

curl http://localhost:11434/api/generate -d '{
  "model": "gemma4",
  "prompt": "Why is the sky blue?",
  "stream": false
}'

Response Payload​

{
  "model": "gemma4",
  "created_at": "2026-06-05T14:46:05Z",
  "response": "The sky is blue because of Rayleigh scattering.",
  "done": true,
  "total_duration": 4820918000,
  "load_duration": 12093000,
  "prompt_eval_count": 12,
  "prompt_eval_duration": 18239000,
  "eval_count": 210,
  "eval_duration": 4790290000
}

3. Structured Output (JSON Schema)​

curl -X POST http://localhost:11434/api/generate -H "Content-Type: application/json" -d '{
  "model": "gemma4",
  "prompt": "GoingMerry is 22 years old. Respond in JSON.",
  "stream": false,
  "format": {
    "type": "object",
    "properties": {
      "age": { "type": "integer" }
    },
    "required": ["age"]
  }
}'

Response Payload​

{
  "model": "gemma4",
  "created_at": "2026-06-05T14:46:05Z",
  "response": "{\n  \"age\": 22\n}",
  "done": true
}