Skip to main content
Glama

ByteBot MCP Server

Production-grade Model Context Protocol (MCP) server for ByteBot's dual-API architecture, providing intelligent hybrid workflow orchestration for autonomous task execution and desktop computer control.

Overview

This MCP server integrates ByteBot's Agent API (task management) and Desktop API (computer control) into a unified interface for AI assistants like Claude. It enables:

  • Autonomous Task Execution: Create and manage tasks for ByteBot to execute independently

  • Direct Computer Control: Mouse, keyboard, screen capture, and file operations

  • Hybrid Workflows: Intelligent orchestration with automatic monitoring and intervention handling

  • Real-time Updates: Optional WebSocket support for live task status notifications

Features

Agent API Tools (Task Management)

  • bytebot_create_task - Create new tasks with priority levels

  • bytebot_list_tasks - List and filter tasks by status/priority

  • bytebot_get_task - Get detailed task information with message history

  • bytebot_get_in_progress_task - Check currently running task

  • bytebot_update_task - Update task status or priority

  • bytebot_delete_task - Delete tasks

Desktop API Tools (Computer Control)

Mouse Operations:

  • bytebot_move_mouse - Move cursor to coordinates

  • bytebot_click - Click with left/right/middle button

  • bytebot_drag - Drag from one position to another

  • bytebot_scroll - Scroll in any direction

Keyboard Operations:

  • bytebot_type_text - Type text strings

  • bytebot_paste_text - Paste text (for special characters)

  • bytebot_press_keys - Keyboard shortcuts (Ctrl+C, Alt+Tab, etc.)

Screen Operations:

  • bytebot_screenshot - Capture screen as base64 PNG

  • bytebot_cursor_position - Get current cursor position

File I/O:

  • bytebot_read_file - Read file content (base64)

  • bytebot_write_file - Write file content (base64)

System:

  • bytebot_switch_application - Switch to application

  • bytebot_wait - Wait for specified duration

Hybrid Orchestration Tools (Priority 1)

  • bytebot_create_and_monitor_task - Create task and wait for completion

  • bytebot_monitor_task - Monitor existing task until terminal state

  • bytebot_intervene_in_task - Provide help when task needs intervention

  • bytebot_execute_workflow - Multi-step workflow with automatic error recovery

Prerequisites

  • Node.js: 20.x or higher

  • ByteBot Instance: Running and accessible at configured endpoints

    • Agent API (default: http://localhost:9991)

    • Desktop API (default: http://localhost:9990)

Installation

# Clone or download this repository cd bytebot-mcp-server # Install dependencies npm install # Build TypeScript code npm run build

Configuration

1. Create Environment File

Copy the example environment file and customize:

cp .env.example .env

2. Edit .env File

# ByteBot Agent API (Task Management) BYTEBOT_AGENT_URL=http://localhost:9991 # ByteBot Desktop API (Computer Control) BYTEBOT_DESKTOP_URL=http://localhost:9990 # WebSocket Configuration (Optional) BYTEBOT_WS_URL=ws://localhost:9991 ENABLE_WEBSOCKET=false # Server Configuration MCP_SERVER_NAME=bytebot-mcp # Timeouts (milliseconds) REQUEST_TIMEOUT=30000 DESKTOP_ACTION_TIMEOUT=10000 # Retry Configuration MAX_RETRIES=3 RETRY_DELAY=1000 # Monitoring Configuration TASK_POLL_INTERVAL=2000 TASK_MONITOR_TIMEOUT=300000 # File Configuration MAX_FILE_SIZE=10485760 # Logging LOG_LEVEL=info

3. Remote ByteBot Configuration

If ByteBot is running on a remote server:

BYTEBOT_AGENT_URL=http://your-server.com:9991 BYTEBOT_DESKTOP_URL=http://your-server.com:9990 BYTEBOT_WS_URL=ws://your-server.com:9991

MCP Client Setup

Claude Desktop

Add to your Claude Desktop configuration file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json

{ "mcpServers": { "bytebot": { "command": "node", "args": ["/absolute/path/to/bytebot-mcp-server/dist/index.js"], "env": { "BYTEBOT_AGENT_URL": "http://localhost:9991", "BYTEBOT_DESKTOP_URL": "http://localhost:9990" } } } }

Zed Editor

Add to your Zed settings:

{ "context_servers": { "bytebot": { "command": { "path": "node", "args": ["/absolute/path/to/bytebot-mcp-server/dist/index.js"] }, "env": { "BYTEBOT_AGENT_URL": "http://localhost:9991", "BYTEBOT_DESKTOP_URL": "http://localhost:9990" } } } }

Continue.dev

Add to .continue/config.json:

{ "mcpServers": [ { "name": "bytebot", "command": "node", "args": ["/absolute/path/to/bytebot-mcp-server/dist/index.js"], "env": { "BYTEBOT_AGENT_URL": "http://localhost:9991", "BYTEBOT_DESKTOP_URL": "http://localhost:9990" } } ] }

Usage Examples

Example 1: Basic Task Creation

User: Create a task for ByteBot to search Wikipedia for "quantum computing" Claude uses: bytebot_create_task { "description": "Go to wikipedia.org and search for 'quantum computing'", "priority": "MEDIUM" } Response: { "id": "task-123", "status": "PENDING", "priority": "MEDIUM", "createdAt": "2024-01-15T10:30:00Z" }

Example 2: Hybrid Workflow (Create → Monitor → Complete)

User: Create a task to log into example.com and wait for it to complete Claude uses: bytebot_create_and_monitor_task { "description": "Navigate to example.com and log in with credentials from keychain", "timeout": 60000, "pollInterval": 2000 } Response: { "taskId": "task-456", "finalStatus": "COMPLETED", "completedAt": "2024-01-15T10:31:45Z", "messagesCount": 12, "task": { ... full task details ... } }

Example 3: Task Needs Intervention

User: Create a task to fill out a complex form Claude uses: bytebot_create_and_monitor_task { "description": "Fill out the registration form at example.com/register" } Response (after monitoring): { "taskId": "task-789", "finalStatus": "NEEDS_HELP", "task": { "id": "task-789", "status": "NEEDS_HELP", "messages": [ { "role": "assistant", "content": "I need the user's phone number to complete this form" } ] } } User: My phone number is 555-1234 Claude uses: bytebot_intervene_in_task { "taskId": "task-789", "message": "User's phone number is 555-1234", "action": "resume", "continueMonitoring": true } Response: { "taskId": "task-789", "status": "COMPLETED", "intervention": "applied" }

Example 4: Interactive Desktop Control

User: Take a screenshot and click at position (500, 300) Claude uses: bytebot_screenshot Response: { "screenshot": "iVBORw0KG..." } Claude uses: bytebot_click { "x": 500, "y": 300, "button": "left" } Response: ✓ bytebot_click completed successfully

Example 5: Multi-Step Workflow

User: Execute a workflow to open Firefox, navigate to GitHub, and take a screenshot Claude uses: bytebot_execute_workflow { "steps": [ { "name": "Open Firefox", "description": "Switch to Firefox browser application" }, { "name": "Navigate to GitHub", "description": "Navigate to github.com in the browser" }, { "name": "Take Screenshot", "description": "Capture a screenshot of the GitHub homepage" } ], "priority": "HIGH" } Response: { "steps": [ { "name": "Open Firefox", "taskId": "task-001", "status": "COMPLETED" }, { "name": "Navigate to GitHub", "taskId": "task-002", "status": "COMPLETED" }, { "name": "Take Screenshot", "taskId": "task-003", "status": "COMPLETED" } ], "overallStatus": "completed", "totalInterventions": 0 }

Example 6: File Operations

User: Read the contents of /home/user/data.txt Claude uses: bytebot_read_file { "path": "/home/user/data.txt" } Response: { "content": "SGVsbG8gV29ybGQh..." } // Base64 encoded

Troubleshooting

Error: "Cannot connect to ByteBot server"

Cause: ByteBot is not running or endpoint URL is incorrect

Solution:

  1. Verify ByteBot is running: curl http://localhost:9991/tasks

  2. Check .env file has correct URLs

  3. Ensure no firewall blocking connections

Error: "Request to ByteBot timed out"

Cause: Task took longer than configured timeout

Solution:

  1. Increase REQUEST_TIMEOUT in .env for Agent API calls

  2. Increase DESKTOP_ACTION_TIMEOUT for Desktop API calls

  3. Use bytebot_create_and_monitor_task with custom timeout:

    { "description": "Long running task", "timeout": 600000 }

Error: "Task with ID xyz not found"

Cause: Task was deleted or ID is incorrect

Solution:

  1. List all tasks: bytebot_list_tasks

  2. Verify task ID from response

  3. Check if task was accidentally deleted

Warning: "Screenshot size is 8.5MB"

Cause: Screenshot is very large (high resolution display)

Solution:

  1. This is just a warning, screenshot still works

  2. Consider reducing screen resolution if frequently capturing screenshots

  3. Screenshots >5MB will show this warning

Error: "Task must be in NEEDS_HELP state"

Cause: Attempting to intervene in task that doesn't need help

Solution:

  1. Check task status first: bytebot_get_task

  2. Only use bytebot_intervene_in_task when status is NEEDS_HELP

  3. Use bytebot_update_task to manually change status if needed

WebSocket Connection Failed

Cause: WebSocket URL incorrect or ByteBot doesn't support WebSocket

Solution:

  1. Set ENABLE_WEBSOCKET=false in .env to disable WebSocket

  2. Server will automatically fall back to HTTP polling

  3. WebSocket is optional - all features work without it

Error: "File size exceeds maximum allowed size"

Cause: Trying to upload/read file larger than 10MB

Solution:

  1. Increase MAX_FILE_SIZE in .env (in bytes)

  2. Split large files into smaller chunks

  3. Compress files before uploading

API Reference

Task Priority Levels

  • LOW - Background tasks, non-urgent

  • MEDIUM - Default priority (recommended)

  • HIGH - Important tasks, process soon

  • URGENT - Critical tasks, process immediately

Task Lifecycle States

  1. PENDING - Task created, waiting to start

  2. IN_PROGRESS - Task currently executing

  3. NEEDS_HELP - Task blocked, requires intervention

  4. NEEDS_REVIEW - Task complete but needs verification

  5. COMPLETED - Task finished successfully

  6. CANCELLED - Task cancelled by user

  7. FAILED - Task failed with error

Mouse Buttons

  • left - Primary button (default)

  • right - Context menu button

  • middle - Scroll wheel click

Scroll Directions

  • up - Scroll up

  • down - Scroll down

  • left - Scroll left

  • right - Scroll right

Common Applications

  • firefox - Mozilla Firefox

  • chrome - Google Chrome

  • safari - Safari (macOS)

  • terminal - Terminal/Command Prompt

  • vscode - Visual Studio Code

Architecture

┌─────────────────────────────────────────────┐ │ MCP Client (Claude) │ └─────────────────┬───────────────────────────┘ │ stdio transport ┌─────────────────▼───────────────────────────┐ │ ByteBot MCP Server │ │ ┌────────────────────────────────────────┐ │ │ │ Agent Tools │ Desktop Tools │ │ │ │ Hybrid Orchestrator │ │ │ └────────────┬──────────────┬─────────────┘ │ └───────────────┼──────────────┼───────────────┘ │ │ ┌──────────▼──┐ ┌──────▼──────┐ │ Agent API │ │ Desktop API │ │ (port 9991) │ │ (port 9990) │ └─────────────┘ └─────────────┘ │ │ ┌──────▼───────────────────▼──────┐ │ ByteBot Instance │ └─────────────────────────────────┘

Development

Build

npm run build

Type Check

npm run type-check

Watch Mode

npm run dev

Environment Variables Reference

Variable

Default

Description

BYTEBOT_AGENT_URL

http://localhost:9991

ByteBot Agent API endpoint

BYTEBOT_DESKTOP_URL

http://localhost:9990

ByteBot Desktop API endpoint

BYTEBOT_WS_URL

ws://localhost:9991

WebSocket endpoint for real-time updates

ENABLE_WEBSOCKET

false

Enable WebSocket connections

MCP_SERVER_NAME

bytebot-mcp

Server identifier

REQUEST_TIMEOUT

30000

HTTP request timeout (ms)

DESKTOP_ACTION_TIMEOUT

10000

Desktop action timeout (ms)

MAX_RETRIES

3

Maximum retry attempts for failed requests

RETRY_DELAY

1000

Initial retry delay (ms)

TASK_POLL_INTERVAL

2000

Task status polling interval (ms)

TASK_MONITOR_TIMEOUT

300000

Maximum task monitoring duration (ms)

MAX_FILE_SIZE

10485760

Maximum file size in bytes (10MB)

LOG_LEVEL

info

Logging level (debug/info/warn/error)

License

MIT

Support

For issues and questions:

Version History

1.0.0 (2024-01-15)

  • Initial release

  • Agent API integration (task management)

  • Desktop API integration (computer control)

  • Hybrid orchestration tools

  • WebSocket support for real-time updates

  • Comprehensive error handling and retry logic

  • Full TypeScript implementation with strict typing

-
security - not tested
F
license - not found
-
quality - not tested

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/sensuslab/spark-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server