什么是 Bilt?
让 AI agents 通过自然语言工作流自主构建并部署 mobile apps,减少手工开发步骤。
README
Bilt MCP Server
<div align="center">
Enable AI agents to autonomously build and deploy full-stack mobile applications
Documentation • Sign Up • Examples • Support
</div>What is Bilt MCP Server?
Bilt's Model Context Protocol (MCP) server gives AI agents 8 powerful tools to autonomously create, build, and deploy production-ready mobile applications through natural language.
Key Features
- 🤖 Agent-First Design - Built specifically for AI autonomy
- 🚀 Zero Setup - Remote SSE server, no local installation
- 🏗️ Complete Lifecycle - From project creation to production deployment
- ⚡ Real-Time Updates - SSE transport for live progress tracking
- 🔒 Secure - Bearer token authentication, rate limiting
- 📱 Mobile Focus - Optimized for mobile app development
Quick Start
1. Get API Token
Sign up at bilt.me/sign-up and generate an API key.
2. Configure Your MCP Client
<details> <summary><b>Claude Desktop</b></summary>Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS):
{
"mcpServers": {
"bilt": {
"transport": {
"type": "sse",
"url": "https://mcp.bilt.me/mcp/sse",
"headers": {
"Authorization": "Bearer bilt_live_YOUR_TOKEN_HERE"
}
}
}
}
}
Restart Claude Desktop.
</details> <details> <summary><b>OpenClaw</b></summary>Edit ~/.openclaw/openclaw.json:
{
"mcpServers": {
"bilt": {
"transport": {
"type": "sse",
"url": "https://mcp.bilt.me/mcp/sse",
"headers": {
"Authorization": "Bearer bilt_live_YOUR_TOKEN_HERE"
}
}
}
}
}
Restart: openclaw gateway restart
Create .cursor/mcp_config.json in your project:
{
"mcpServers": {
"bilt": {
"transport": {
"type": "sse",
"url": "https://mcp.bilt.me/mcp/sse",
"headers": {
"Authorization": "Bearer bilt_live_YOUR_TOKEN_HERE"
}
}
}
}
}
Reload Cursor window: Cmd/Ctrl+Shift+P → "Reload Window"
3. Test It
Ask your AI agent:
"Use Bilt to create a todo app with authentication"
Watch it autonomously create, build, and deploy a production app! 🎉
Available Tools
The Bilt MCP server provides 8 tools for complete application lifecycle management:
| Tool | Purpose |
|---|---|
bilt_list_projects | List all projects for authenticated user |
bilt_get_project | Get detailed project information |
bilt_create_project | Create a new project |
bilt_get_session | Get current workflow session status |
bilt_send_message | Execute build/deploy instructions |
bilt_resume_workflow | Resume paused workflows |
bilt_cancel_workflow | Cancel running workflows |
bilt_get_messages | Retrieve workflow message history |
Example Workflow
Agent: bilt_create_project("todo-app")
→ Returns: { id: "proj_abc123" }
Agent: bilt_get_session()
→ Returns: { session_id: "sess_xyz789" }
Agent: bilt_send_message(
session_id: "sess_xyz789",
message: "Create a todo list with add, delete, and complete features"
)
→ Builds app...
Agent: bilt_send_message(
session_id: "sess_xyz789",
message: "Deploy to production"
)
→ Returns: { url: "https://todo-app-abc123.bilt.app" }
Documentation
- Introduction - What is Bilt MCP?
- Quick Start - 5-minute setup guide
- API Reference - Complete tool documentation
- Integration Guides - Client-specific setup
- Examples - Step-by-step workflows
Why MCP?
Model Context Protocol is the emerging standard for AI tool integration. Benefits:
✅ Universal Compatibility - Works with any MCP-compatible client ✅ Standardized Interface - No custom integrations needed ✅ Auto-Discovery - Agents find and use tools automatically ✅ Session Management - Built-in state handling ✅ Real-Time Updates - SSE transport for progress tracking
Technical Details
Connection Details
- Server URL:
https://mcp.bilt.me/mcp/sse - Transport: SSE (Server-Sent Events)
- Alternative: HTTP at
https://mcp.bilt.me/mcp - Authentication: Bearer token (
bilt_live_...) - Rate Limit: 100 requests/minute
Requirements
- MCP-compatible client (Claude Desktop, OpenClaw, Cursor, etc.)
- Bilt account with API token
- Internet connection (remote server)
Supported Platforms
- ✅ macOS
- ✅ Windows
- ✅ Linux
- ✅ Any platform with MCP client support
Use Cases
🤖 Autonomous App Development
AI agents build entire applications from scratch - from initial requirements to production deployment.
⚡ Rapid Prototyping
Generate working prototypes in minutes for testing ideas and gathering feedback.
🔧 Feature Development
Add new features to existing projects through natural language instructions.
👥 Multi-Agent Collaboration
Multiple agents work on different aspects of a project simultaneously.
Examples
See the examples/ directory for:
- Complete workflow examples
- Multi-step builds
- Error handling patterns
- Advanced use cases
FAQ
<details> <summary><b>Is this open source?</b></summary>The Bilt MCP server is a proprietary service. This repository contains documentation and integration examples. The server itself runs on Bilt's infrastructure.
</details> <details> <summary><b>How much does it cost?</b></summary>Bilt offers a free tier for experimentation. See bilt.me/pricing for details.
</details> <details> <summary><b>Can I self-host?</b></summary>Currently, Bilt MCP is cloud-only. We may offer self-hosted options for enterprise customers in the future.
</details> <details> <summary><b>What kind of apps can I build?</b></summary>Bilt specializes in mobile applications built with React Native and Expo. Supports iOS and Android.
</details> <details> <summary><b>Is my data secure?</b></summary>Yes. All communication uses HTTPS. API tokens are required for authentication. Your code and data remain private.
</details>Support
-
📚 Documentation: bilt.me/docs
-
💬 Discord: Join our community
-
📧 Email: support@bilt.me
-
🐛 Issues: GitHub Issues
Community
License
Copyright © 2024-2026 Bilt. All rights reserved.
The Bilt MCP server is proprietary software. This repository contains documentation and examples under the MIT License. See LICENSE for details.
Acknowledgments
Built with ❤️ by the Bilt team.
MCP specification by Anthropic.
<div align="center">
Made with Model Context Protocol
</div>常见问题
Bilt 是什么?
让 AI agents 通过自然语言工作流自主构建并部署 mobile apps,减少手工开发步骤。
相关 Skills
Claude接口
by anthropics
面向接入 Claude API、Anthropic SDK 或 Agent SDK 的开发场景,自动识别项目语言并给出对应示例与默认配置,快速搭建 LLM 应用。
✎ 想把Claude能力接进应用或智能体,用claude-api上手快、兼容Anthropic与Agent SDK,集成路径清晰又省心
RAG架构师
by alirezarezvani
聚焦生产级RAG系统设计与优化,覆盖文档切块、检索链路、索引构建、召回评估等关键环节,适合搭建可扩展、高准确率的知识库问答与检索增强应用。
✎ 面向RAG落地,把知识库、向量检索和生成链路系统串联起来,做架构设计时更清晰,也更少踩坑。
计算机视觉
by alirezarezvani
聚焦目标检测、图像分割与视觉系统落地,覆盖 YOLO、DETR、Mask R-CNN、SAM 等方案,适合定制数据集训练、推理优化及 ONNX/TensorRT 部署。
✎ 把目标检测、图像分割到推理部署串成完整工程链路,主流框架与 YOLO、DETR、SAM 等方案都覆盖,落地视觉 AI 会省心很多。
相关 MCP Server
顺序思维
编辑精选by Anthropic
Sequential Thinking 是让 AI 通过动态思维链解决复杂问题的参考服务器。
✎ 这个服务器展示了如何让 Claude 像人类一样逐步推理,适合开发者学习 MCP 的思维链实现。但注意它只是个参考示例,别指望直接用在生产环境里。
知识图谱记忆
编辑精选by Anthropic
Memory 是一个基于本地知识图谱的持久化记忆系统,让 AI 记住长期上下文。
✎ 帮 AI 和智能体补上“记不住”的短板,用本地知识图谱沉淀长期上下文,连续对话更聪明,数据也更可控。
PraisonAI
编辑精选by mervinpraison
PraisonAI 是一个支持自反思和多 LLM 的低代码 AI 智能体框架。
✎ 如果你需要快速搭建一个能 24/7 运行的 AI 智能体团队来处理复杂任务(比如自动研究或代码生成),PraisonAI 的低代码设计和多平台集成(如 Telegram)让它上手极快。但作为非官方项目,它的生态成熟度可能不如 LangChain 等主流框架,适合愿意尝鲜的开发者。