• Orchestrate agent tooling with Strata, our progressive MCP router that keeps function calling under control.
• Run more than one hundred first-party MCP servers (OAuth-ready where it matters) straight from the repository.
• Instrument and monitor every deployment by embedding the new MCP Handle Status service (adapted from an open-source dashboard by JustSong) for rich health and analytics insights.
• Ship clients quickly through ready-to-run Discord, Slack, WhatsApp, CLI, and web MCP clients.

• Strata MCP router – dynamic orchestration so agents only see the tools they need when they need them (open-strata/).
• MCP server catalog – >100 integrations spanning SaaS, data, comms, and infra (mcp_servers/).
• Multi-channel MCP clients – production-grade bots powered by the official MCP SDK (mcp-clients/).
• MCP Handle Status – embedded status center with Go/React backend+frontend (mcp_handle_status/).
• Docs & guides – comprehensive Mintlify docs rewired for the new brand (docs/).

workspace/
├── docs/                    # Mintlify documentation, API reference, and tutorials
├── mcp_servers/             # First-party MCP server implementations
├── mcp-clients/             # Discord, Slack, Web, WhatsApp MCP clients
├── open-strata/             # Progressive MCP orchestration engine
├── mcp_handle_status/       # New status & analytics dashboard (Go + React)
├── scripts/                 # Helper scripts (e.g., start_status_service.sh)
└── NOTICE, LICENSE          # Apache-2.0 root license + third-party attributions

1. Install prerequisites

   • Python 3.12+
   • Node.js 18+
   • Go 1.20+
   • Docker (optional but recommended)

2. Clone the repository

   git clone https://github.com/WeatherPal-AI/MCP-handle.git
   cd MCP-handle

3. Install dependencies (optional selective steps)

   # Install Strata CLI
   cd open-strata
   uv pip install -r requirements.txt  # or your preferred tool
   cd ..
   
   # Bootstrap client dependencies (example: Discord bot)
   cd mcp-clients
   uv sync
   cd ..

4. Launch Strata locally

   cd open-strata
   uv run strata --help

5. Run the MCP Handle Status dashboard

   ./scripts/start_status_service.sh
   # Frontend dev server : http://localhost:5173
   # API & dashboard     : http://localhost:3000

   The helper script creates a default .env, builds the React frontend, and boots the Go backend. Logs are written to mcp_handle_status/backend.log.

• Every server lives under mcp_servers/<integration> with language-specific tooling.
• OAuth-capable servers consume the _oauth_support/ helpers and respect the new MCP_HANDLE_* environment variables.
• Docker images and package metadata now point to the WeatherPal-AI namespace (ghcr.io/weatherpal-ai/..., @weatherpal-ai/...).
• Example recipes in examples/ map directly to the new Python (mcp_handle) and TypeScript packages.

# Example: run the GitHub MCP server in watch mode
cd mcp_servers/github
go run ./...

The integrated status service replaces the proprietary dashboards from the original project.

• Live health checks for stdio, SSE, and streamable HTTP servers.
• Installation manager with npm/PyPI bootstrap flows.
• Analytics (usage, latency, failure modes) backed by SQLite out-of-the-box.

Key commands:

# Build the dashboard container
cd mcp_handle_status
docker build -t weatherpal-ai/mcp-handle-status .

# Run natively (already done by the helper script)
./run.sh

• Fork, branch, and submit PRs targeting main.
• Keep code comments concise and add tests where behaviour changes.
• If you extend third-party code (e.g., the MCP Handle Status component), preserve upstream attribution in NOTICE and the component’s local LICENSE.

• Root license: Apache-2.0 (see LICENSE).
• Embedded code: MIT-licensed portions sourced from an open project by JustSong remain under MIT inside mcp_handle_status/.
• Maintainer: WeatherPal-AI（低温宝）

Open issues and roadmap items are tracked in the GitHub repository. Join the discussion, file feature requests, and help the MCP ecosystem stay open and agent-friendly.