Skip to content
MuvomiMuvomiMuvomi Docs

Connect AI assistants

Muvomi works with AI assistants that support the Model Context Protocol (MCP), such as ChatGPT, Claude, Claude Code, Cursor, VS Code, Gemini CLI and Windsurf. Once connected, you can ask your assistant to make a film from a Muvomi template without opening the studio. It can choose a template, write the words, apply your website’s brand, set the length and pace, record a voiceover, render the film and hand you the MP4.

The assistant uses the same tools as the studio, so a film made this way is checked, voiced and rendered exactly like one you make by hand. It’s saved to your account and appears in your library.

The server address

Section titled “The server address”

Give your assistant this address:

https://muvomi.com/api/mcp

The AI apps section of your account page shows the same address and lists the apps you’ve connected.

Add Muvomi to your assistant

Section titled “Add Muvomi to your assistant”

Add a custom MCP server and enter https://muvomi.com/api/mcp as its address.

Sign in and allow access

Section titled “Sign in and allow access”

  1. When your assistant connects for the first time, it opens Muvomi in your browser.

  2. Sign in to your Muvomi account if you aren’t already.

  3. A consent page asks Let [app] work in your Muvomi? It shows the account you’re signed in as and what the app will be able to do: see your saved films and make new ones from templates, record voiceovers (which use your credits), render films (free), and see your email and plan. Check the address it says you’ll return to, then choose Allow, or Don’t allow.

  4. Go back to your assistant. It can now work with your Muvomi account.

From then on the assistant works as you. The films it makes are saved to your account and appear in Projects as Motion film, and the files it adds go to your private storage.

What you can ask for

Section titled “What you can ask for”

Talk to your assistant in plain words, for example:

  • “Show me Muvomi’s templates for a product launch.”
  • “Make a 15-second film from the Feature Spotlight template for our new app, using the brand from example.com.”
  • “Make it the extended length with a relaxed pace, then record the voiceover.”
  • “Render it and give me the MP4.”

Behind the scenes, your assistant uses these tools:

Tool What it does
list_templates, get_template Lists the templates, and each one’s fields, how much text fits, and its lengths, paces and voices
create_film, update_film Makes a film from a template and sets its words, format, length and pace
get_film, list_films Reads a film back with its script and timings, and lists your films
apply_brand Takes a website’s logo, name and colour, and what the site says, to write the film from
set_film_media Puts a logo or picture into the film from a public image address (up to 25 MB)
record_voice Records the voiceover from the film’s own words and fits the film to it
render_film, get_render Starts the MP4 and follows it to the finished file

A finished render is handed to you as a download link. Open it in a browser where you’re signed in to Muvomi to download the MP4.

Cost

Section titled “Cost”

Recording a voiceover uses the voiceover price, 20 credits for every minute of speech started, and is given back if no recording is made. Making and changing films, reading a website and rendering are free. Free-plan renders carry the watermark, as they do in the studio.

Disconnect an assistant

Section titled “Disconnect an assistant”

Open your account page, find AI apps, and press Disconnect beside the app. Its access ends at once. You can connect it again at any time.

Templates come in three lengths (standard, extended and full) and three paces: dense (as designed), steady (about 20% longer) and relaxed (about 45% longer).

Limits

Section titled “Limits”
  • Most of the 52 templates can be used from an assistant. A few that are written as a directed storyboard can’t be used this way yet. Ask your assistant to list templates to see which are available.
  • A voiceover is available for templates that have a voice option.
  • Each account can read one website and record one voiceover at a time, and up to 5 exports can wait to render.
  • Files on your own computer can’t be used; give the assistant a public image address instead.
  • An assistant needs to sign in through a browser. Scripts and automation tools without a browser can’t connect yet.