Sponsored by Deepsite.site

Buildable

Created By
chunkydotdev8 months ago
Buildable generates a phased backlog, scaffolds a private repo with CI, then serves JSON tasks through the Model-Controller-Prompt protocol. Any MCP-compatible assistant can pull tasks, push code, and move Kanban cards to Done.
Content

🌟 What is Buildable?

Buildable (bldbl.dev) is an AI-powered development platform that makes any project buildable. It provides:

  • AI-Generated Build Plans: Comprehensive project roadmaps with implementation details
  • Smart Task Management: Automated task breakdown with dependencies and priorities
  • AI Assistant Integration: Direct integration with Claude, GPT, and other AI assistants
  • Real-time Collaboration: Seamless human-AI collaboration on complex projects
  • Progress Tracking: Live monitoring of development progress and blockers

🚀 Features

  • Full Project Integration: Get complete project context, plans, and task details
  • Autonomous Task Management: Start, update progress, and complete tasks
  • Human Collaboration: Create discussions for questions and blockers
  • Real-time Progress Tracking: Live updates and status monitoring
  • Type-Safe API: Full TypeScript support with comprehensive type definitions
  • Claude Desktop Ready: CLI interface for seamless Claude Desktop integration

📦 Installation

npm install @bldbl/mcp

🚀 Quick Start

Step 1: Install the MCP Server

npm install -g @bldbl/mcp

Step 2: Get Your Buildable API Key

  1. Go to bldbl.dev and create an account
  2. Create or select a project
  3. Go to Project SettingsAI AssistantsGenerate API Key

Step 3: Configure Your AI Assistant

🤖 For Claude Desktop:

Add this to your Claude Desktop config file (~/.config/claude/claude_desktop_config.json):

{
  "mcpServers": {
    "buildable": {
      "command": "npx",
      "args": ["-y", "@bldbl/mcp"],
      "env": {
        "BUILDABLE_API_KEY": "bp_your_api_key_here",
        "BUILDABLE_PROJECT_ID": "your-project-id",
        "BUILDABLE_AI_ASSISTANT_ID": "claude-desktop"
      }
    }
  }
}

Minimal setup - only API URL is auto-detected.

✨ For Cursor AI:

Option 1: One-Click Setup (Recommended)

  1. Go to your project on bldbl.dev
  2. Navigate to Integrations
  3. Click "Add to Cursor" - Done! 🎉

Option 2: Manual Configuration Add this to your Cursor AI config file:

{
  "mcpServers": {
    "buildable": {
      "command": "npx",
      "args": ["-y", "@bldbl/mcp"],
      "env": {
        "BUILDABLE_API_KEY": "bp_your_api_key_here",
        "BUILDABLE_PROJECT_ID": "your-project-id",
        "BUILDABLE_AI_ASSISTANT_ID": "cursor-ide"
      }
    }
  }
}

Minimal setup - only API URL is auto-detected.

Step 4: Start AI-Powered Development!

Once configured, your AI assistant has access to these powerful tools:

  • 🏗️ get_project_context - Understand your entire project structure
  • 📋 get_next_task - Find what to work on next based on priorities
  • ▶️ start_task - Begin working on a specific task
  • 📊 update_progress - Report detailed progress updates
  • complete_task - Mark tasks finished with comprehensive notes
  • 💬 create_discussion - Ask questions when you need human input
  • 🔍 health_check - Verify connection to Buildable

💡 Example AI Conversation:

You: "What should I work on next in my project?"

AI: *calls get_project_context & get_next_task*
    "I see you're building a React dashboard app. The next 
     high-priority task is 'Implement user authentication'. 
     Should I start working on it?"

You: "Yes, go ahead!"

AI: *calls start_task*
    "Perfect! I'm starting the authentication task. I'll use 
     NextAuth.js with GitHub OAuth. Let me break this down:
     1. Install dependencies
     2. Set up NextAuth config  
     3. Create login/logout components
     4. Add protected routes
     
     I'll update you on progress as I work!"

[AI works and calls update_progress periodically]

AI: *calls update_progress*
    "Progress update: 75% complete! I've implemented:
     ✅ NextAuth.js configuration
     ✅ GitHub OAuth provider setup  
     ✅ Login/logout components
     🔄 Currently working on: Protected route middleware
     
     Should I continue with the remaining work?"

🔧 Advanced Configuration

Environment Variables

3 environment variables are recommended for proper AI assistant tracking:

export BUILDABLE_API_KEY="bp_your_api_key_here"           # Your API key (required)
export BUILDABLE_PROJECT_ID="your-project-id"             # Target project (required)
export BUILDABLE_AI_ASSISTANT_ID="my-ai-assistant"        # AI Assistant ID (recommended)

# Optional (has smart default):
# export BUILDABLE_API_URL="https://bldbl.dev/api"        # API endpoint (default)

Why AI Assistant ID matters: This helps you track which AI assistant is doing what work in your project dashboard.

CLI Usage

You can also run the MCP server directly:

# Install globally and use bldbl command
npm install -g @bldbl/mcp
bldbl

# Or run with npx (no installation needed)
npx @bldbl/mcp

Multiple Projects

To work with multiple Buildable projects, create different MCP server configs:

{
  "mcpServers": {
    "buildable-frontend": {
      "command": "npx",
      "args": ["-y", "@bldbl/mcp"],
      "env": {
        "BUILDABLE_API_KEY": "bp_frontend_key_here",
        "BUILDABLE_PROJECT_ID": "frontend-project-id",
        "BUILDABLE_AI_ASSISTANT_ID": "claude-frontend"
      }
    },
    "buildable-backend": {
      "command": "npx", 
      "args": ["-y", "@bldbl/mcp"],
      "env": {
        "BUILDABLE_API_KEY": "bp_backend_key_here",
        "BUILDABLE_PROJECT_ID": "backend-project-id",
        "BUILDABLE_AI_ASSISTANT_ID": "claude-backend"
      }
    }
  }
}

🛠️ API Reference

BuildPlannerMCPClient

The main client class for interacting with Buildable projects.

Constructor

new BuildPlannerMCPClient(config: BuildPlannerConfig, options?: ClientOptions)

Config Parameters:

  • apiUrl: Buildable API URL (defaults to 'https://bldbl.dev/api')
  • apiKey: Your Buildable API key (starts with 'bp_')
  • projectId: Target project ID
  • aiAssistantId: Unique identifier for your AI assistant
  • timeout: Request timeout in milliseconds (default: 30000)

Options:

  • retryAttempts: Number of retry attempts (default: 3)
  • retryDelay: Delay between retries in ms (default: 1000)

Methods

getProjectContext(): Promise<ProjectContext>

Get complete project context including plan, tasks, and recent activity.

getNextTask(): Promise<NextTaskResponse>

Get the next recommended task to work on based on dependencies and priority.

startTask(taskId: string, options?: StartTaskOptions): Promise<StartTaskResponse>

Start working on a specific task with optional approach and timing estimates.

updateProgress(taskId: string, progress: ProgressUpdate): Promise<ProgressResponse>

Update progress on the current task with detailed status information.

completeTask(taskId: string, completion: CompleteTaskRequest): Promise<CompleteTaskResponse>

Mark a task as completed with detailed completion information.

createDiscussion(discussion: CreateDiscussionRequest): Promise<DiscussionResponse>

Create a discussion/question for human input when you need guidance.

healthCheck(): Promise<{status: string, timestamp: string}>

Check connectivity and health of the Buildable API.

disconnect(): Promise<void>

Properly disconnect and cleanup the client connection.

🔐 Authentication

  1. Generate API Key: Go to your Buildable project → AI Assistant tab → Generate API Key
  2. Secure Storage: Store your API key securely (environment variables recommended)
  3. Key Format: API keys start with bp_ followed by project and random identifiers

🐛 Error Handling

The client includes comprehensive error handling:

try {
  const context = await client.getProjectContext();
} catch (error) {
  if (error.code === 'UNAUTHORIZED') {
    console.error('Invalid or expired API key');
  } else if (error.code === 'PROJECT_NOT_FOUND') {
    console.error('Project not found or access denied');
  } else {
    console.error('API error:', error.message);
  }
}

🔄 Development Workflow

Typical AI assistant workflow with Buildable:

  1. Initialize - Connect to Buildable with API key
  2. Get Context - Understand the project structure and current state
  3. Find Work - Get the next priority task
  4. Start Task - Begin working with approach and estimates
  5. Progress Updates - Regular progress reports with details
  6. Ask Questions - Create discussions for blockers or decisions
  7. Complete Task - Finish with comprehensive completion notes
  8. Repeat - Continue with next tasks

📊 Usage Statistics

// Get usage statistics for your AI assistant
const stats = await client.getUsageStats();
console.log(`Tasks completed: ${stats.tasksCompleted}`);
console.log(`Average completion time: ${stats.avgCompletionTime}min`);
console.log(`Success rate: ${stats.successRate}%`);

⚡ CLI Usage

Once installed, you can use the CLI in several ways:

# Run directly with npx (no installation needed)
npx @bldbl/mcp

# Or install globally and use the bldbl command
npm install -g @bldbl/mcp
bldbl

# For Claude Desktop, use the bldbl command in your config

Environment Variables Required:

  • BUILDABLE_API_URL - Your Buildable API URL
  • BUILDABLE_API_KEY - Your API key (starts with 'bp_')
  • BUILDABLE_PROJECT_ID - Target project ID
  • BUILDABLE_AI_ASSISTANT_ID - Unique assistant identifier

🧪 Testing

The package includes comprehensive test utilities:

import { createTestClient } from '@bldbl/mcp/test';

// Create a test client with mock responses
const testClient = createTestClient({
  mockProject: {
    id: 'test-project',
    title: 'Test Project'
  }
});

// Use in your tests
await testClient.startTask('test-task-id');

🏗️ Built With

  • TypeScript - Type-safe development
  • Model Context Protocol (MCP) - Standardized AI assistant communication
  • Node.js - Runtime environment
  • REST API - Simple and reliable communication

📄 License

Copyright © 2025 Buildable Team. All rights reserved.

This software is proprietary and confidential. Unauthorized copying, distribution, or use is strictly prohibited.


Made with ❤️ by the Buildable team

Buildable is a commercial AI-powered development platform. Visit bldbl.dev to get started.

🆘 Support


Built with ❤️ by the BuildPlanner team

Server Config

{
  "mcpServers": {
    "buildable": {
      "command": "npx",
      "args": [
        "-y",
        "@bldbl/mcp"
      ],
      "env": {
        "BUILDABLE_API_KEY": "<YOUR-API-TOKEN>",
        "BUILDABLE_PROJECT_ID": "<YOUR-PROJECT-ID>",
        "BUILDABLE_AI_ASSISTANT_ID": "<NAME-YOUR-ASSISTANT>"
      }
    }
  }
}
Recommend Servers
TraeBuild with Free GPT-4.1 & Claude 3.7. Fully MCP-Ready.
EdgeOne Pages MCPAn MCP service designed for deploying HTML content to EdgeOne Pages and obtaining an accessible public URL.
TimeA Model Context Protocol server that provides time and timezone conversion capabilities. This server enables LLMs to get current time information and perform timezone conversions using IANA timezone names, with automatic system timezone detection.
DeepChatYour AI Partner on Desktop
WindsurfThe new purpose-built IDE to harness magic
Jina AI MCP ToolsA Model Context Protocol (MCP) server that integrates with Jina AI Search Foundation APIs.
BlenderBlenderMCP connects Blender to Claude AI through the Model Context Protocol (MCP), allowing Claude to directly interact with and control Blender. This integration enables prompt assisted 3D modeling, scene creation, and manipulation.
Serper MCP ServerA Serper MCP Server
ChatWiseThe second fastest AI chatbot™
MiniMax MCPOfficial MiniMax Model Context Protocol (MCP) server that enables interaction with powerful Text to Speech, image generation and video generation APIs.
Tavily Mcp
Howtocook Mcp基于Anduin2017 / HowToCook (程序员在家做饭指南)的mcp server,帮你推荐菜谱、规划膳食,解决“今天吃什么“的世纪难题; Based on Anduin2017/HowToCook (Programmer's Guide to Cooking at Home), MCP Server helps you recommend recipes, plan meals, and solve the century old problem of "what to eat today"
Visual Studio Code - Open Source ("Code - OSS")Visual Studio Code
AiimagemultistyleA Model Context Protocol (MCP) server for image generation and manipulation using fal.ai's Stable Diffusion model.
Amap Maps高德地图官方 MCP Server
Y GuiA web-based graphical interface for AI chat interactions with support for multiple AI models and MCP (Model Context Protocol) servers.
MCP AdvisorMCP Advisor & Installation - Use the right MCP server for your needs
Zhipu Web SearchZhipu Web Search MCP Server is a search engine specifically designed for large models. It integrates four search engines, allowing users to flexibly compare and switch between them. Building upon the web crawling and ranking capabilities of traditional search engines, it enhances intent recognition capabilities, returning results more suitable for large model processing (such as webpage titles, URLs, summaries, site names, site icons, etc.). This helps AI applications achieve "dynamic knowledge acquisition" and "precise scenario adaptation" capabilities.
CursorThe AI Code Editor
Baidu Map百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright McpPlaywright MCP server