Skip to main content

Generate Chat Completion

Generate the next message in a conversation thread with a specified model.

POST /api/chat

Request Parameters​

  • model (string, required): Model tag.
  • messages (array of objects, required): Conversation message thread history. Each object requires:
    • role (string): system, user, assistant, or tool.
    • content (string): Message content.
    • images (array of strings, optional): Base64-encoded images.
    • tool_calls (array of objects, optional): Tool function calls returned by the model.
  • stream (boolean, default true): Return streamed response chunks.
  • format (string/object, optional): Set to "json" or provide a JSON Schema layout.
  • tools (array of objects, optional): Array of function tools the model can call.
  • options (object, optional): Override Modelfile parameters.
  • keep_alive (string/number, default "5m"): Cache duration in memory.

Examples​

1. Simple Chat Request (Streaming)​

curl http://localhost:11434/api/chat -d '{
  "model": "gemma4",
  "messages": [
    { "role": "user", "content": "Hello!" }
  ]
}'

Chunk Payload​

{
  "model": "gemma4",
  "created_at": "2026-06-05T14:46:10Z",
  "message": {
    "role": "assistant",
    "content": "Hello"
  },
  "done": false
}

2. Multi-turn Conversation (Non-Streaming)​

curl http://localhost:11434/api/chat -d '{
  "model": "gemma4",
  "stream": false,
  "messages": [
    { "role": "user", "content": "What is the capital of France?" },
    { "role": "assistant", "content": "The capital of France is Paris." },
    { "role": "user", "content": "What is its population?" }
  ]
}'

Response Payload​

{
  "model": "gemma4",
  "created_at": "2026-06-05T14:46:12Z",
  "message": {
    "role": "assistant",
    "content": "The population of Paris is approximately 2.1 million within the city limits."
  },
  "done": true,
  "total_duration": 1928010000,
  "load_duration": 4810000,
  "prompt_eval_count": 28,
  "prompt_eval_duration": 11928000,
  "eval_count": 85,
  "eval_duration": 1910800000
}