A new release, llm-typesafe 0.1a0, expands the LLM tool with support for TypeSafe AI's Jev model. Install the plugin in your llm environment with:
llm install llm-typesafe
After installation, set your TypeSafe API key using the following command and paste the key when prompted:
llm keys set typesafe
Paste key
The article notes that TypeSafe issues API keys via a waitlist, but that the wait appears to move relatively quickly.
Usage examples
The plugin demonstrates multiple question types with command-line examples.
- Yes/no ("noul") questions: the example below checks whether a message explicitly requests a refund:
llm -m jev 'Please refund my last payment.'
-s 'Does this message explicitly request a refund?'
The output is returned as JSON, for example:
{"type": "noul", "noul": 0.99}
- Choice questions: you can provide categories and criteria for routing a message. Example:
cat message.txt |
llm -m jev
-s 'Which team should handle this message? If billing and technical issues both occur, choose billing.'
-o answer_type choice
-o criteria '{
"billing":"Charges, invoices, payments, or refunds",
"technical":"Problems installing or using the product",
"other":"Neither category fits"
}'
- Scoring questions: to rate how reproducible a reported problem is, use a scoring query as in this example:
cat report.txt |
llm -m jev
-s 'How reproducible is the problem described in this report?'
-o answer_type score
-o criteria '[
"No reproduction instructions",
"Some instructions, but important steps are missing",
"Complete steps with expected and actual results"
]'
For further configuration options and more examples, consult the README which contains additional usage details.
Version and context
The release is titled llm-typesafe 0.1a0. The plugin aims to make the Jev model accessible from the LLM command-line environment, supporting multiple output formats (noul, choice, score).
