personal_mcp

Personal MCP server — v0: personal_ping only. Learn stdio in Cursor, publish docs on GitHub Pages, run HTTP on Render.

Two deploy surfaces (same repo, different jobs)

Related MCP server: romate-mcp-server

Phases

Local setup

cd personal_mcp
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env

Cursor (stdio)

.cursor/mcp.json: venv Python + absolute server.py. No secrets. Reload MCP → call personal_ping.

Test locally

.venv/bin/python -c "import asyncio; from server import mcp; print(asyncio.run(mcp.call_tool('personal_ping', {})).structured_content)"

HTTP smoke test (optional)

export MCP_TRANSPORT=streamable-http MCP_API_KEY=dev-secret MCP_PORT=8008
python server.py &
curl -sS http://127.0.0.1:8008/health
curl -sS -o /dev/null -w "%{http_code}\n" http://127.0.0.1:8008/mcp   # expect 401 without key
curl -sS -o /dev/null -w "%{http_code}\n" -H "X-API-Key: dev-secret" http://127.0.0.1:8008/mcp
kill %1

GitHub Pages

1. Repo → Settings → Pages → Build and deployment → Source: GitHub Actions.

2. Push to main (workflow .github/workflows/deploy-pages.yml uploads docs/).

3. Site: https://adhimulambhargavsaiviswanath-05.github.io/personal_mcp/

(Branch /docs deploy still works, but Actions is recommended for the static index.html site.)

Render

1. Merge v0 to main (Render blueprint uses branch: main).

2. Render Dashboard → New → Blueprint → connect personal_mcp (or Web Service with same settings as render.yaml).

3. Env (blueprint sets these): MCP_TRANSPORT=streamable-http, MCP_API_KEY (generated), PYTHON_VERSION=3.12.8. Render injects PORT automatically.

4. After deploy: copy service URL into docs/index.md, enable Pages if not already.

5. curl https://YOUR-SERVICE.onrender.com/health

MCP endpoint path: /mcp (FastMCP default). Protect with X-API-Key from Render env.

Git workflow

• feature/* → PR → dev → PR → main for releases / deploy

Security

Never commit .env. Store MCP_API_KEY and future tokens only in Render env or local .env.