Works with Muse
Your docs
as a file for Muse
Muse works with the files that you upload in a conversation. Download your docs site as one text file, upload it to Muse and ask your questions.
- Publishes from GitHub on every push
- Works with Mintlify repositories
- API console from your OpenAPI file
- MCP server and llms.txt for AI agents
- Ask AI and support tickets
- Analytics for people and AI agents
Give Muse your docs as a file
Muse is Meta's personal AI agent, in the Muse app and on the web. Meta offers it in the US and Canada. Other AI apps read a Revdoku docs site through its MCP server. MCP is a standard way for AI apps to use outside tools. Meta's help pages describe no way to add an MCP server to Muse. So give Muse your docs as a file.
- Open the
llms-full.txtfile of the docs site, for examplehttps://docs.example.com/llms-full.txt. This file holds every page of the site. - Save the file on your phone or computer.
- Start a conversation in Muse and upload the file.
- Ask your question about the docs. Tell Muse to use only the uploaded file.
For one page only, save the Markdown copy of that page instead. Each page links its Markdown copy in the For AI agents footer. Muse reads PDF, DOCX, plain text, Markdown, XLSX and CSV files, and PNG, JPEG, GIF and WebP images.
The file starts each page with its title, so Muse can name the page that it used. A file is a copy of the docs at one time. After you publish a change, download the file again and upload the new copy.
Docs work that Muse can do with the file
- A new developer asks how to send the first API request. Muse writes the steps from the docs and names the page of each step.
- Before a release, you upload the file of last month and the file of today. Muse lists the pages that changed.
- Muse turns the API reference into a CSV file with one row for each endpoint.
Muse saves the files that it makes in the Library tab and shares them in the conversation.
A Muse prompt for your docs
Upload llms-full.txt to one conversation. Then send this message:
I uploaded the docs of the Acme API as llms-full.txt.
Write the steps that a new developer must do to send the first request.
Use only the uploaded file. After each step, write the title of its page.
If the file does not answer a step, write "missing". Do not guess.Custom connectors, approvals and data in Muse
Muse connects to the services in Settings → Connectors. If a service is not in that list, you can ask Muse to create a custom connector. Muse guides you through the process, which can include the API information of the service.
Muse keeps the credentials of a custom connector in its Secure Credentials Store. Meta does not review custom connectors.
By default, Muse asks for your approval before important actions, such as sending an email. You can change when Muse asks in Settings. To stop a connector, disconnect it in Settings → Connectors.
Muse stores the files that you upload in its system files. The setting Help improve our AI models in Settings → Data controls is on when you first use Muse. Turn it off before you upload private documents.
An AI app that can add MCP servers can also read your support tickets and publish your docs. Examples are ChatGPT and Claude.
How Muse works with Connectors ·
How to manage your Muse data · Docs for AI agents · Revdoku API