Skip to content

Latest commit

 

History

History
225 lines (169 loc) · 2.61 KB

API.md

File metadata and controls

225 lines (169 loc) · 2.61 KB

API Reference

Base URL: http://localhost:8000

Interactive docs: http://localhost:8000/docs

Tasks

Execute Task

POST /tasks

Request body:

{
  "command": "open Chrome and search for weather",
  "device_id": "DEVICE_SERIAL",
  "model": "qwen2.5-vl:7b"
}

device_id and model are optional. Defaults to first available device and configured model.

Response:

{
  "task_id": "uuid",
  "status": "completed",
  "command": "open Chrome and search for weather",
  "actions": [
    {
      "step": 0,
      "action": {"action_type": "tap", "x": 540, "y": 1200},
      "result": {"status": "success"}
    }
  ],
  "error": null
}

Get Task

GET /tasks/{task_id}

Returns task details and execution history.

Devices

List Devices

GET /devices

Response:

[
  {
    "device_id": "R5CRA1XXXXX",
    "model": "Pixel 7",
    "status": "online",
    "connection_type": "usb"
  }
]

Connect WiFi Device

POST /devices/connect

Request body:

{
  "address": "192.168.1.100",
  "port": 5555
}

Disconnect Device

POST /devices/{device_id}/disconnect

Screen Stream

WebSocket Stream

WS /stream/{device_id}

Receives JSON frames:

{
  "type": "frame",
  "data": "base64_encoded_png",
  "device_id": "DEVICE_SERIAL"
}

Frames are sent approximately every 500ms.

Scenarios

List Scenarios

GET /scenarios

Create Scenario

POST /scenarios

Request body:

{
  "name": "morning_routine",
  "description": "Check weather and calendar",
  "steps": [
    {"action": "open", "target": "Weather", "wait": 2.0},
    {"action": "tap", "target": "today"},
    {"action": "open", "target": "Calendar"}
  ]
}

Get / Update / Delete Scenario

GET    /scenarios/{name}
PUT    /scenarios/{name}
DELETE /scenarios/{name}

Run Scenario

POST /scenarios/{name}/run

Voice

Voice Command

POST /voice/command
Content-Type: multipart/form-data

Upload audio file. Returns transcribed text and task result.

Response:

{
  "text": "open settings",
  "task_id": "uuid",
  "status": "completed"
}

Voice Status

GET /voice/status

Response:

{
  "stt_available": true,
  "tts_available": true,
  "wake_word_active": false
}

Models

List Models

GET /models

Response:

[
  {"name": "qwen2.5-vl:7b", "size": "4.7GB", "active": true},
  {"name": "llava:7b", "size": "4.5GB", "active": false}
]

Switch Model

POST /models/switch

Request body:

{
  "model": "llava:7b"
}