llm-typesafe 0.1a0
Simon Willison has released llm-typesafe 0.1a0, a plugin for LLM that adds support for TypeSafe AI's new Jev model.
The release is installed with `llm install llm-typesafe`, followed by setting an API key via `llm keys set typesafe`. Willison notes that a key can be obtained from TypeSafe AI and that the waitlist "seems to move pretty fast."
The plugin supports three kinds of questions. Yes/no "noul" questions are submitted with the `-m jev` model flag, a question passed with `-s`, and return output such as `{"type": "noul", "noul": 0.99}`. An example given is asking whether the message "Please refund my last payment." explicitly requests a refund.
Choice questions accept an `answer_type` of `choice` with a criteria object naming options and their descriptions, such as billing, technical, and other. An example asks which team should handle a message, with instructions that billing takes precedence when both billing and technical issues occur.
Scoring questions accept an `answer_type` of `score` with a criteria array of ordered descriptions, including a report example about how reproducible a described problem is.
llm-typesafe 0.1a0 is tagged under projects, llm, and jev. More details are available in the README.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
Release: llm-typesafe 0.1a0 I built this new plugin for LLM to add support for TypeSafe AI's new Jev model . Install it like this: llm install llm-typesafe Then set an API key ( get one here , the waitlist seems to move pretty fast): llm keys set typesafe # Paste key And now you can ask yes/no "noul" questions like this: llm -m jev 'Please refund my last payment.' \ -s 'Does this message explicitly request a refund?' Output: {"type": "noul", "noul": 0.99} Or choice questions like this: cat message.txt | llm -m jev