Get Fast Yes or No, Sorting, and Rating Decisions Through Jev and Your Muse

Through a custom connector, your muse can use Jev to get a fast, scored decision: yes or no with a probability, one pick from a list you define, or a rating on a scale you set.

What it does

Jev is a decision model, not a chat model. You give it the facts of a case, called the state, plus typed questions, and it returns numbers in about 300 milliseconds. It never writes paragraphs. Connected this way, the muse gets 6 tools. One answers yes or no with a probability. One sorts a case into one option from a set you define, with a probability per option and a confidence score. One rates a case on ordered levels, from 2 to 10. One asks several questions about the same facts in one call, so the facts go once. The set also includes a free playbook on phrasing questions, and a tool listing the models your key can use. This suits sorting work: is this message urgent, which team owns this request, is this a duplicate.

How to use it

1. Start an AnythingMCP workspace and install the Jev by TypeSafe connector.
2. Create a key in the TypeSafe console and paste it where the connector asks for TYPESAFE_API_KEY.
3. In Muse, open Settings, then Connectors, and add a custom connector with the MCP server address AnythingMCP gives you. Sign in and approve.
4. Ask the muse to read the Jev playbook first, then set a threshold: act on confident answers, bring unsure cases back to you.
5. Give the muse a sorting job in plain words, with the options or levels written out, for example sorting requests into three named queues.

What you need

• A Muse account that can add custom connectors
• An AnythingMCP account with the Jev by TypeSafe connector added
• A TypeSafe account and API key

What to watch for

• A probability near 0.5 means the model is unsure, not a middle answer. Set your threshold so unsure cases come back to a person.
• Confidence is not a promise the pick is right, so spot-check early batches before you trust a rule.
• Jev reads text only, no pictures, and works best in English. Other languages work, but check confidence more closely.
• Cost is per input token, about 0.04 dollars per million, and output is free. Keep each question's facts tight.
• This is a custom connector, and Meta does not review custom connectors. You can disconnect it in Settings at any time. The muse prepares, you decide.