Lumicat

Connector

Lumicat in your AI assistant

Lumicat ships a Model Context Protocol server, so ChatGPT, Claude and other MCP-compatible assistants can create images, videos, talking avatars, music and speech with the same models and the same credit balance you use in the Lumicat app and Studio.

Connect

The server URL is https://mcp.lumicat.app/mcp. Sign-in happens through your existing Lumicat account (Apple, Google or email); the assistant never sees your password.

ChatGPT

  1. Open Settings → Apps & Connectors.
  2. Search for Lumicat in the directory, or add the URL above under Developer mode.
  3. Click Connect and sign in with your Lumicat account.

Claude

  1. Open Settings → Connectors → Add custom connector.
  2. Paste the URL above and confirm.
  3. Connect and sign in when Claude asks.

Cursor, Claude Code and Codex

  • Cursor: Add to Cursor, or add the URL above under Settings → MCP. Then click Connect next to Lumicat and sign in.
  • Claude Code: run claude mcp add --transport http lumicat https://mcp.lumicat.app/mcp, then /mcp and choose Authenticate.
  • Codex: run codex mcp add lumicat --url https://mcp.lumicat.app/mcp and sign in when it asks (or later with codex mcp login lumicat).

Any other MCP client that supports OAuth 2.1 works the same way: point it at the URL and it asks you to sign in before it connects.

What it can do

  • generate_image — Text-to-image and image-to-image with any Studio image model.
  • generate_video — Text- or image-to-video clips.
  • generate_avatar_video — Talking-avatar videos from a photo and a script.
  • enhance_draft — Re-render a FLUX 3 draft video at 1080p.
  • generate_audio — Music (with your own lyrics) and sound effects, each with a title and cover art.
  • generate_speech / change_voice / change_video_voice — Read text aloud, or re-voice a recording or video in another voice.
  • merge_audio_video — Put music or a voice-over on a video, with an optional start offset.
  • mix_audio — Layer tracks into one at the levels you choose.
  • merge_videos — Join two or more clips into one video, in the order you give them.
  • upload_media — Bring a photo or clip into a generation from a public URL or a file attached in the chat.
  • list_models / get_model / estimate_cost — Browse the catalogue and see the exact credit price before running anything.
  • get_account / list_generations / get_generation — Check your balance, revisit past results and follow a running job.

Results appear inline in the conversation and are saved to your Lumicat gallery, so you can keep editing them in the app or Studio.

Credits & limits

Every generation costs credits, priced per model exactly as in Studio. The assistant is told the price before it runs a job and you can cap what a single request may spend. Credits are never sold inside the chat; see how credits work. To protect your balance, connector usage is limited per hour and per day (the current limits are returned by get_account).

Privacy

When you connect, the assistant receives a short-lived token for your Lumicat account. We process the prompts and media the assistant sends solely to run the generation you asked for, and keep a per-call audit record (tool name, outcome, credits) for a limited time to operate the service and enforce the limits above. We never sell this data or use it to train models. Details are in the “AI assistant connectors” section of our privacy policy. You can disconnect at any time from the assistant's settings; this revokes its access immediately.

Support

Something not working? Contact us and mention the assistant you're using.