jevkit

How it uses Jev

TypeSafe’s Jev returns typed decisions instead of prose. jevkit is a CLI for it: ask Jev questions about a passage, get back probabilities, labels, and scores. jev lint validates a question set offline, before you spend anything on a call.

Highlights

• 🎯 Typed questions : noul (probability), choice (label), score (level), sent in terse YAML or JSON, by file, stdin, --question-set inline, or --noul TEXT for a quick yes/no
• 🦀 Offline linting : 20 rules that catch billed-but-useless questions before a call; documented exit codes for CI
• 🔩 Machine-first I/O : JSON on stdout (compact when piped, pretty on a TTY; --compact / --pretty to force), diagnostics on stderr, 0/1/2 exit codes (ok / input·credential·API fail…
• 📊 Usage ledger : jev ask --log appends one JSON line per call (request, response, tokens, cost) to /.local/state/jev/usage.jsonl
• 🛠 Config : user defaults in /.config/jev/config.toml ( provider , model , log ), layered flags env config defaults
• 🔐 Keyring auth : keys stored in the OS credential store, never argv or shell history; fingerprint display rather than the key itself

Getting started

No file, either: a yes/no question passed inline is a one-liner.