MCP

Auto-configure MCP servers for your stack

MCP (Model Context Protocol) servers give AI coding assistants superpowers — access to databases, API docs, deployment tools, and more. Caliber discovers which MCP servers are relevant for your project and configures them for Claude Code, Cursor, and other tools.

What is MCP?

The Model Context Protocol is an open standard that lets AI assistants connect to external tools and data sources. Instead of the AI guessing about your database schema or API endpoints, MCP servers provide real data directly to the model.

Think of MCP servers as plugins for your AI assistant. A PostgreSQL MCP server lets the AI query your database schema. A GitHub MCP server lets it read issues and PRs. A documentation MCP server gives it access to your internal docs.

How Caliber handles MCP setup

1. 1Scans your project — detects your database, cloud provider, CI/CD, and key dependencies
2. 2Recommends MCP servers — suggests servers that match your stack (e.g., PostgreSQL MCP for a project using Prisma + PostgreSQL)
3. 3Generates configs — writes the correct MCP configuration format for your tool (Claude Code uses a different format than Cursor)
4. 4Preserves settings — if you already have MCP configs with permissions or custom settings, Caliber merges without overwriting

Supported tools

Caliber generates MCP configs in the correct format for each tool:

• Claude Code — writes to .claude/settings.local.json
• Cursor — writes to .cursor/mcp.json
• Any MCP-compatible tool — Windsurf, Cline, and other tools supporting the Model Context Protocol

Get started

Caliber also generates CLAUDE.md and .cursorrules from the same scan — everything your AI coding assistant needs.