Use Gemini Models, Files, and Pictures Through Your Muse
Through an AnythingMCP custom connector, your muse can call Google Gemini models with your own API key: ask questions, read a file from a public link, and create or edit pictures that appear in the chat.
What it does
Gemini is Google's family of AI models. Connected this way, the muse gets 13 tools that run on your key. It can list the models your key can use and read the details of one model, both at no cost. It can ask a model a question and hold a multi-turn conversation. It can answer questions about a file at a public link, such as an image, a PDF, a text or spreadsheet file, or a short video, up to 15 MB. It can count how much text a request contains before you send it, so a long job can be priced first. It can turn text into the numeric form used for search and grouping, in batches of up to twenty. It can create a picture from a description and edit an existing picture from your words, showing the result in the chat.
How to use it
- Start an AnythingMCP account and add the Google Gemini API connector.
- In Google AI Studio, create an API key for a project you control. Enter it only where the connector asks, never in a chat.
- In Muse, open Settings, then Connectors, and add a custom connector with the MCP server address AnythingMCP gives you. Sign in and approve.
- For pictures, ask for the cost and approve it before the muse creates or edits one. For files, paste a public link that opens without a sign-in.
What you need
- A Muse account that can add custom connectors
- An AnythingMCP account with the Google Gemini API connector added
- A Google account and a Gemini API key; picture models need billing turned on for the project
What to watch for
- Almost every call is billed to your key's project or counts against its free limits. The model list and the token count are the free checks, so use them first.
- Pictures have no free tier. The default picture model costs about three cents a picture and the strongest about thirteen cents, and the muse confirms with you before each one. Pictures over 4 MB are not returned.
- A file link must open in public without a login, or the read fails. The connector does not upload files from your computer.
- Speech and video generation, live audio and video, and batch jobs are not covered by this connector.