puter/doc/api/drivers.md

Puter Drivers

POST /drivers/call

Notes

• HTTP response status - A successful driver response, even if the response is an error message, will always have HTTP status 200. Note that sometimes this will include rate limit and usage limit errors as well.

This endpoint allows you to call a Puter driver. Whether or not the driver call fails, this endpoint will respond with HTTP 200 OK. When a driver call fails, you will get a JSON response from the driver with

Parameters

Parameters are provided in the request body. The content type of the request should be application/json.

• interface: string
  • description: The type of driver to call. For example, LLMs use the interface called puter-chat-completion.
• service: string
  • description: The name of the service to use. For example, the claude service might be used for puter-chat-completion.
• method: string
  • description: The name of the method to call. For example, LLMs implement complete which does a chat completion, and list which lists models.
• args: object
  • description: Parametized arguments for the driver call. For example, puter-chat-completion's complete method supports the arguments messages and temperature (and others), so you might set this to { "messages": [...], "temperature": 1.2 }

Example

{
    "interface": "<name of interface>",
    "service": "<name of service>",
    "method": "<name of method>",
    "args": { "parametized": "arguments" }
}

Response

• Error Response - Driver error responses will always have status 200, content type application/json, and a response body in this format:

  {
    "success": false,
    "error": {
        "code": "string identifier for the error",
        "message": "some message about the error",
    }
  }
  

• Success Response - The success response is either a JSON response wrapped in { "success": true, "result": ___ }, or a response with a Content-Type that is not application/json.

  {
    "success": true,
    "result": {}
  }