Enable AstraSync's Identity, Trust and Verification capabilities through Model Context Protocol (MCP) - from account creation to agent registration, verification, and cryptographic signing.
The AstraSync MCP Bridge connects AI assistants and development tools to the AstraSync AI Agent Registry. Through a single MCP integration, you can:
- Create developer accounts - Sign up without leaving your AI assistant
- Manage API credentials - Generate and manage API keys programmatically
- Register AI agents - Add agents to the blockchain-based registry
- Verify agent identity - Check registration status and trust scores
- Generate crypto keypairs - Enable cryptographic signing for ownership proof
Think of it as complete identity infrastructure for AI agents, accessible through natural conversation.
- Universal Protocol: Works with any MCP-compatible platform via HTTP
- Simple Setup: Quick self-hosted deployment with minimal configuration
- Instant Integration: AI assistants can register agents during natural conversation
- Real-time Compliance: Every agent gets a verifiable identity and trust score
# Clone and install
git clone https://github.com/AstraSyncAI/astrasync-mcp-bridge.git
cd astrasync-mcp-bridge
npm install
# Configure environment
cp .env.example .env
# Edit .env if needed (defaults to production API at https://astrasync.ai/api)
# Start the bridge
npm start
# Or for development with auto-reload:
npm run devThe bridge will run on http://localhost:3000 by default.
Once running, configure your AI assistant to use the bridge:
{
"mcpServers": {
"astrasync": {
"url": "http://localhost:3000/mcp/v1",
"transport": "http"
}
}
}Test your setup: Visit http://localhost:3000/mcp/test for an interactive test interface.
The AstraSync MCP Bridge uses the official @modelcontextprotocol/sdk for automatic protocol compliance and future-proof compatibility.
Benefits:
- ✅ Automatic version negotiation - Supports latest MCP protocol (2025-06-18)
- ✅ Built-in transports - StreamableHTTP with SSE support included
- ✅ Standardized error handling - Proper McpError types
- ✅ Reduced maintenance - SDK handles protocol updates automatically
- ✅ Battle-tested - Used by all major MCP implementations
Dependencies:
{
"@modelcontextprotocol/sdk": "^1.21.1",
"express": "^4.21.2",
"cors": "^2.8.5",
"dotenv": "^16.4.7"
}Why SDK over custom implementation?
The SDK saves ~22-30 hours of maintenance over 2 years:
- Custom implementation: ~12-16 hours/year (protocol updates, bug fixes, testing)
- SDK implementation: ~1 hour/year (dependency updates only)
Plus automatic compliance with future MCP spec changes.
Register a new AI agent with AstraSync's compliance registry (requires authentication).
Required parameters:
agentName: Name of the AI agentagentDescription: What the agent doesagentOwner: Name of the agent owner or companyemail: Your AstraSync account email
Authentication (one required):
password: Your account password, ORapiKey: Your AstraSync API key
Note: You must have an AstraSync account to register agents. Use create_account first if needed.
Verify if an agent is registered and compliant.
Required parameters:
agentId: The agent ID to verify
Create a new AstraSync developer account.
Required parameters:
email: Email address for the accountpassword: Password (min 8 characters)fullName: Full name of the developeraccountType:individualorbusiness(optional, defaults toindividual)
Returns: Account creation confirmation with dashboard link
Generate a new API key for your AstraSync account.
Required parameters:
email: Account email addresspassword: Account passwordkeyName: Name/label for this API key
Returns: API key (save it securely - you won't see it again!)
Note: Requires existing AstraSync account (use create_account first)
Generate a crypto keypair for signing agent registrations.
Required parameters:
email: Account email addresspassword: Account passwordkeyName: Name/label for this keypair (optional)
Returns: Public key and confirmation that mnemonic was emailed
Tier limits:
- Free tier: 1 crypto keypair
- Developer tier: Unlimited keypairs
Note: Your mnemonic phrase will be sent to your email. Save it securely offline!
Tested platforms where the AstraSync MCP Bridge works:
| Platform | Compatibility | Integration Method | Status | How to Use |
|---|---|---|---|---|
| Claude Desktop | ✅ Fully Compatible | HTTP MCP Server | Tested | Add to claude_desktop_config.json |
| ChatGPT Desktop | ✅ Fully Compatible | HTTP MCP Server | Tested | Add to MCP settings |
| OpenAI Agents SDK | ✅ Fully Compatible | HTTP MCP Server | Tested | Configure in agent code |
| Google Gemini CLI | ✅ Fully Compatible | Official Extension | Tested | Install extension |
| Cursor | ✅ Fully Compatible | HTTP MCP Server | Tested | Add to MCP configuration |
| Windsurf | ✅ Fully Compatible | HTTP MCP Server | Tested | Add to MCP configuration |
| Cline | ✅ Fully Compatible | HTTP MCP Server | Tested | VS Code extension settings |
| Replit | ✅ Compatible | HTTP MCP Server | Community tested | Configure MCP server |
| Sourcegraph | ✅ Compatible | HTTP MCP Server | Community tested | Configure in settings |
| Letta | ✅ Compatible | HTTP MCP Server | Community tested | Memory-enabled agents |
| Salesforce Agentforce | 🔄 In Progress | MCP Client | Testing | MCP support in pilot |
| ChatGPT Web | Direct API via GPT Actions | Limited | Use JSON-RPC directly | |
| Claude Web | Copy/Paste | Manual | No direct integration |
Legend:
- ✅ Fully Compatible: Officially tested and supported
- ✅ Compatible: Community tested, expected to work
- 🔄 In Progress: Under testing
⚠️ Workaround: Alternative integration methods available
Gemini CLI - Use our official extension:
gemini extensions install https://github.com/AstraSyncAI/astrasync-gemini-extensionClaude Desktop - Add to claude_desktop_config.json:
{
"mcpServers": {
"astrasync": {
"url": "http://localhost:3000/mcp/v1",
"transport": "http"
}
}
}ChatGPT Desktop - Add to MCP settings:
{
"mcpServers": {
"astrasync": {
"url": "http://localhost:3000/mcp/v1",
"transport": "http"
}
}
}Salesforce Agentforce - Available in Pilot (July 2025 release)
For platforms without MCP support, you can call the bridge directly:
curl -X POST http://localhost:3000/mcp/v1 \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "register_agent",
"arguments": {
"agentName": "Your Agent",
"agentDescription": "What it does",
"agentOwner": "Your Company",
"email": "[email protected]",
"password": "your-password"
}
}
}'Note: For production use, use apiKey instead of password for better security.
{
"jsonrpc": "2.0",
"id": 1,
"result": {
"content": [{
"type": "text",
"text": "Successfully registered agent: Customer Support Bot\nAgent ID: ASTRAS-789456\nTrust Score: 45/100 (Free tier baseline)\nBlockchain Status: Queued for verification\n\nView agent card: https://astrasync.ai/agents/ASTRAS-789456"
}]
}
}{
"jsonrpc": "2.0",
"id": 2,
"result": {
"content": [{
"type": "text",
"text": "✓ Account created successfully!\nEmail: [email protected]\nType: individual\n\nYou can now:\n• Generate API keys with generate_api_key\n• Create crypto keypairs with create_crypto_keypair (Developer tier)\n• Register agents with your account\n\nLogin to your dashboard: https://astrasync.ai/dashboard"
}]
}
}AI Assistant → MCP Protocol (HTTP) → AstraSync MCP Bridge → AstraSync API
PORT: Server port (default: 3000)NODE_ENV: Environment mode (development/production)ASTRASYNC_API_URL: AstraSync API endpoint (default: https://astrasync.ai/api)
Docker:
docker build -t astrasync-mcp-bridge .
docker run -p 3000:3000 -e ASTRASYNC_API_URL=https://astrasync.ai/api astrasync-mcp-bridgeCloud Platforms:
- Works on any Node.js hosting (Railway, Render, Fly.io, AWS, GCP, Azure)
- Set
ASTRASYNC_API_URLenvironment variable if needed - Ensure port 3000 (or custom PORT) is accessible
npm run dev # Auto-reload on file changes
npm start # Production mode- Platform Integration Guide - Detailed setup for each platform
- API Reference - Technical protocol details
- Troubleshooting - Common issues and solutions
- Gemini CLI Extension - Native extension for Google's Gemini CLI (Production)
- Salesforce App - Agentforce integration for Salesforce (Coming Soon)
- AWS Marketplace - Enterprise deployment on AWS (Coming Soon)
- Node.js SDK - Universal SDK with auto-detection for 5+ agent formats
- Python SDK - Python integration for agent registration
- Main API - RESTful API for direct integration
AstraSync is building the trust infrastructure for AI agents. Learn more at astrasync.ai.
- Platform: astrasync.ai
- Documentation: astrasync.ai/docs
- GitHub: github.com/AstraSyncAI
MIT License - see LICENSE for details.
Making AI agents accountable, one registration at a time.