API Reference

Responses

Copy Markdown

Open in Claude

Open in ChatGPT

Open in Cursor

---

Copy Markdown

View as Markdown

Create a new response

POST/v1/responses

Create a model response. Supports text conversations and computer-use (CUA) workflows. Set stream=true for server-sent events. Include tools with type "computer_use" for CUA mode, which returns structured computer_call actions. Use previous_response_id to chain multi-turn conversations.

Body ParametersJSONExpand Collapse

input: optional array of object { call_id, content, output, 2 more }

call_id: optional string

content: optional array of ContentBlock { image_url, text, type }

image_url: optional string

text: optional string

type: optional "input_text" or "output_text" or "summary_text" or "input_image"

One of the following:

"input_text"

"output_text"

"summary_text"

"input_image"

output: optional ContentBlock { image_url, text, type }

image_url: optional string

text: optional string

type: optional "input_text" or "output_text" or "summary_text" or "input_image"

One of the following:

"input_text"

"output_text"

"summary_text"

"input_image"

role: optional string

type: optional "computer_call_output"

instructions: optional string

max_output_tokens: optional number

model: optional string

previous_response_id: optional string

stream: optional boolean

temperature: optional number

tools: optional array of object { display_height, display_width, environment, type }

display_height: optional number

display_width: optional number

environment: optional string

type: optional "computer_use"

top_p: optional number

ReturnsExpand Collapse

ResponsesResponse = object { id, computer_id, created_at, 9 more }

id: optional string

computer_id: optional string

created_at: optional string

max_output_tokens: optional number

model: optional string

object: optional string

output: optional array of V2GoBackendInternalServiceOutputItem { id, action, call_id, 5 more }

id: optional string

action: optional V2GoBackendInternalServiceAction { button, end_x, end_y, 10 more }

button: optional string

end_x: optional number

end_y: optional number

keys: optional array of string

result: optional string

scroll_x: optional number

scroll_y: optional number

status: optional string

text: optional string

type: optional "click" or "double_click" or "triple_click" or 14 more

One of the following:

"click"

"double_click"

"triple_click"

"right_click"

"type"

"key"

"keypress"

"key_down"

"key_up"

"scroll"

"hscroll"

"navigate"

"drag"

"wait"

"terminate"

"answer"

"done"

url: optional string

x: optional number

y: optional number

call_id: optional string

content: optional array of ContentBlock { image_url, text, type }

image_url: optional string

text: optional string

type: optional "input_text" or "output_text" or "summary_text" or "input_image"

One of the following:

"input_text"

"output_text"

"summary_text"

"input_image"

role: optional string

status: optional string

summary: optional array of ContentBlock { image_url, text, type }

image_url: optional string

text: optional string

type: optional "input_text" or "output_text" or "summary_text" or "input_image"

One of the following:

"input_text"

"output_text"

"summary_text"

"input_image"

type: optional "reasoning" or "computer_call" or "computer_call_output" or "message"

One of the following:

"reasoning"

"computer_call"

"computer_call_output"

"message"

status: optional string

temperature: optional number

top_p: optional number

usage: optional V2GoBackendInternalServiceResponsesUsage { input_tokens, output_tokens, total_tokens }

input_tokens: optional number

output_tokens: optional number

total_tokens: optional number

viewport: optional V2GoBackendInternalServiceViewport { environment, height, width }

environment: optional string

height: optional number

width: optional number

Create a new response

HTTP

HTTP

TypeScript

Python

curl https://api.tzafon.ai/v1/responses \
    -H 'Content-Type: application/json' \
    -H "Authorization: Bearer $TZAFON_API_KEY" \
    -d '{}'

200 example

{
  "id": "id",
  "computer_id": "computer_id",
  "created_at": "created_at",
  "max_output_tokens": 0,
  "model": "model",
  "object": "object",
  "output": [
    {
      "id": "id",
      "action": {
        "button": "button",
        "end_x": 0,
        "end_y": 0,
        "keys": [
          "string"
        ],
        "result": "result",
        "scroll_x": 0,
        "scroll_y": 0,
        "status": "status",
        "text": "text",
        "type": "click",
        "url": "url",
        "x": 0,
        "y": 0
      },
      "call_id": "call_id",
      "content": [
        {
          "image_url": "image_url",
          "text": "text",
          "type": "input_text"
        }
      ],
      "role": "role",
      "status": "status",
      "summary": [
        {
          "image_url": "image_url",
          "text": "text",
          "type": "input_text"
        }
      ],
      "type": "reasoning"
    }
  ],
  "status": "status",
  "temperature": 0,
  "top_p": 0,
  "usage": {
    "input_tokens": 0,
    "output_tokens": 0,
    "total_tokens": 0
  },
  "viewport": {
    "environment": "environment",
    "height": 0,
    "width": 0
  }
}

Returns Examples

200 example

{
  "id": "id",
  "computer_id": "computer_id",
  "created_at": "created_at",
  "max_output_tokens": 0,
  "model": "model",
  "object": "object",
  "output": [
    {
      "id": "id",
      "action": {
        "button": "button",
        "end_x": 0,
        "end_y": 0,
        "keys": [
          "string"
        ],
        "result": "result",
        "scroll_x": 0,
        "scroll_y": 0,
        "status": "status",
        "text": "text",
        "type": "click",
        "url": "url",
        "x": 0,
        "y": 0
      },
      "call_id": "call_id",
      "content": [
        {
          "image_url": "image_url",
          "text": "text",
          "type": "input_text"
        }
      ],
      "role": "role",
      "status": "status",
      "summary": [
        {
          "image_url": "image_url",
          "text": "text",
          "type": "input_text"
        }
      ],
      "type": "reasoning"
    }
  ],
  "status": "status",
  "temperature": 0,
  "top_p": 0,
  "usage": {
    "input_tokens": 0,
    "output_tokens": 0,
    "total_tokens": 0
  },
  "viewport": {
    "environment": "environment",
    "height": 0,
    "width": 0
  }
}