Vercel MCP Server for Microsoft Copilot Studio
Enables managing Vercel projects, deployments, build logs, domains, and environment variables through natural language in Microsoft Copilot Studio, using streamable HTTP with API key authentication.
README
Vercel MCP Server for Microsoft Copilot Studio
A zero-dependency Model Context Protocol server for the Vercel platform, built specifically to work with Microsoft Copilot Studio (which requires the Streamable HTTP transport and supports API-key authentication).
Why this exists: Vercel's official MCP server (
https://mcp.vercel.com) uses OAuth with dynamic client registration, which Copilot Studio's MCP onboarding cannot complete. This server bridges the gap: the API key you enter in Copilot Studio is your Vercel access token, forwarded per-request to the Vercel REST API. The server itself is stateless and stores nothing.
Tools
| Tool | Description |
|---|---|
get_current_user |
Verify the connection / identify the token owner |
list_teams |
List Vercel teams |
list_projects |
List projects (id, framework, production URL) |
get_project |
Project details incl. git repo link |
list_deployments |
Recent deployments, filter by project/state/target |
get_deployment |
Deployment details incl. error info |
get_deployment_build_logs |
Build log output — great for diagnosing failed builds |
cancel_deployment |
Cancel a building/queued deployment |
promote_deployment |
Promote (roll forward/back) a deployment to production |
list_project_domains |
Custom domains + verification status |
list_env_vars |
Env vars (secret values hidden) |
create_env_var |
Create/upsert an env var |
delete_env_var |
Delete an env var |
Deploy
Option A — Deploy to Vercel (recommended, ~1 minute)
Or with the CLI:
vercel deploy --prod
Your MCP endpoint will be: https://<your-deployment>.vercel.app/mcp
Option B — Self-host (any Node 18+ host)
node server.js # listens on PORT (default 3000), endpoint /mcp
No npm install needed — there are no dependencies.
Connect from Copilot Studio
- Create a Vercel access token at https://vercel.com/account/settings/tokens (scope it to the team/projects you want the agent to manage).
- In Copilot Studio, open your agent → Tools → Add a tool → Model Context Protocol (MCP).
- Fill in:
- Name:
Vercel - Description:
Manage Vercel projects, deployments, build logs, domains and environment variables. - Server URL:
https://<your-deployment>.vercel.app/mcp - Authentication: API key
- Header name / location:
x-api-key(header) - Value: your Vercel access token
- Header name / location:
- Name:
- Click Add → Copilot Studio performs the MCP handshake and lists all 13 tools.
- Create connection, enable it in the connection manager, then test in the Preview pane, e.g.:
- "List my Vercel projects"
- "Show me the build logs for the latest failed deployment of my-site"
- "Add an env var API_URL=https://api.example.com to my-site for preview only"
The server also accepts the token via Authorization: Bearer <token> or x-vercel-token headers, and falls back to a VERCEL_TOKEN environment variable for single-tenant self-hosted deployments (in that case choose No authentication in Copilot Studio — only do this if the server URL itself is protected/private).
Architecture
Copilot Studio ──POST /mcp (JSON-RPC, Streamable HTTP)──▶ this server ──REST──▶ api.vercel.com
x-api-key: <vercel access token> (stateless, no storage)
- Transport: MCP Streamable HTTP, stateless mode (protocol versions 2024-11-05 → 2025-06-18). No sessions, no SSE required — ideal for serverless.
- Auth: pass-through. The server never stores tokens; every request carries the caller's own Vercel token.
- Zero dependencies: plain Node.js ≥18. Nothing to install, no supply chain.
Development & tests
node test/mcp.test.js
The test suite boots the real HTTP server, mocks api.vercel.com, and drives the full MCP handshake (initialize → initialized → tools/list → tools/call), plus error paths (bad token, unknown tool/method, malformed JSON, batching).
Security notes
- Use a scoped Vercel token, ideally on a dedicated team with least privilege.
- Secret env var values are never returned by
list_env_vars. - Deploy under HTTPS only (automatic on Vercel). Copilot Studio requires HTTPS.
- Consider Vercel Deployment Protection bypass headers if you protect the deployment.
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.