MCP Server

Lazer ships a built-in Model Context Protocol (MCP) server that exposes every project operation — projects, scenes, shots, assets, characters, and more — as structured tools that AI agents can call directly.

What is MCP?

The Model Context Protocol is an open standard for connecting AI assistants to external tools and data. Instead of copying and pasting between your AI chat and Lazer, MCP lets the assistant manage your production directly.

Capabilities

The Lazer MCP server provides:

• 53 tools across 11 domains — full CRUD for every object in the hierarchy
• 7 resources — live schema, project trees, and documentation
• 5 prompts — AI workflow templates for scene breakdown, shot planning, and asset review

Architecture

The MCP server runs embedded inside Next.js — no separate process to manage. When you start npm run dev, the MCP HTTP server starts automatically on port 3100.

Client (Claude, ChatGPT, Cursor)
    │
    ▼
Cloudflare Tunnel / localhost
    │
    ▼
Next.js  ─── /mcp route ──→  Embedded MCP Server (port 3100)
                                    │
                                    ▼
                              Prisma / Supabase DB

Two transports are supported:

Transport	Use Case	Auth
STDIO	Local clients (Claude Desktop, Cursor)	--token flag or LAZER_TOKEN env var
Streamable HTTP	Remote clients (ChatGPT, web agents)	OAuth 2.0 or Bearer token

Quick Start

1. ChatGPT — Add as MCP server with OAuth. See Setup Guide.
2. Claude Desktop — Add STDIO config to claude_desktop_config.json. See Setup Guide.
3. Cursor — Add STDIO config to MCP settings. See Setup Guide.

Next Steps

• Setup Guide — Connect your AI client
• Authentication — OAuth, API keys, and token management
• Tools Reference — All 53 tools documented
• Resources & Prompts — Schema access and workflow templates