Developers · MCP Server

Video as a tool call.

Connect the GLX Studio MCP server and your agent can research a prospect, brief a render, and hand back a finished, on-brand MP4 with voiceover in 75+ languages. Same API, same templates, same credits as the REST endpoints. 1 credit = 1 video.

Get an API key OpenAPI 3.1 file

Install in one line

Replace YOUR_KEY with an API key from the Developer Dashboard.

Claude Code

claude mcp add --transport http glx MCP_SERVER_URL --header "Authorization: Bearer YOUR_KEY"

Claude.ai

Settings, Connectors, Add custom connector. Name GLX Studio, URL MCP_SERVER_URL, then paste the key.

Cursor and VS Code

Add an HTTP MCP server named glx with the same URL and an Authorization: Bearer YOUR_KEY header. Both clients also read the official MCP Registry entry com.glxstudio/video.

OpenAI Responses API

{"type":"mcp","server_label":"glx","server_url":"MCP_SERVER_URL","headers":{"Authorization":"Bearer YOUR_KEY"}}

ChatGPT

Build a GPT Action from the OpenAPI file; no server needed.

What the agent does with it

  1. Create a context store. A PDF, deck, brief or prompt. GLX indexes it so the script is written from your content.
  2. Pick a template. Narrative structure, brand treatment, typed variables (account, product line, language).
  3. Render and collect. One async job, 3 to 8 minutes, then a hosted MP4 and thumbnail the agent can send anywhere.

Full workflow for agents: api-agent-guide.md. Reference: developer docs.

Proof

A Fortune 500 technology company ran a 50,000-video personalized campaign on this API in three weeks: conversion 1.2% to 4.8%, 10.4% click-through, under $1 per video at enterprise scale.

Limits worth knowing