Model Context Protocol moderate risk

Vercel AI Docs

An MCP server providing AI-powered search and querying of Vercel AI SDK documentation using the Google Gemini model, enabling contextualized responses.

Connections & Capabilities

Connects To

GitHubVercel

Capabilities

read

Quickstart

Config

{
  "mcpServers": {
    "vercel-ai-docs": {
      "command": "node",  
      "args": ["ABSOLUTE_PATH_TO_PROJECT/dist/main.js"],
      "env": {
        "GOOGLE_GENERATIVE_AI_API_KEY": "your-google-api-key-here"
      }
    }
  }
}

Exposed MCP Tools (3)

moderate

agent-query

Queries the Vercel AI SDK documentation using an AI agent to synthesize information.

AI agent can generate inaccurate or misleading responses; relies on external API.

safe

direct-query

Performs a direct similarity search against the Vercel AI SDK documentation index.

Read-only access to documentation; no side effects.

moderate

clear-memory

Clears the conversation memory for a specific session or all sessions.

Potentially impacts user experience; no direct security risk.

Safety Assessment

This server is relatively safe for querying documentation. The primary risks stem from reliance on the Google Gemini API and the potential for the AI agent to generate incorrect information. Ensure proper API key management and validate agent responses.

Read-only access to documentation via direct query.
Session management provides some context control.
Uses Google Gemini, a managed AI service.
No direct filesystem access.

Relies on a third-party API (Google Gemini) which could introduce vulnerabilities.
API key stored as an environment variable.
AI agent can synthesize information, potentially leading to inaccurate or misleading responses.
Session data is stored, potentially raising privacy concerns if not handled correctly.