Neovim plugin that splits the buffer into functions with Treesitter, has Jev score each one against a plain-language question, and lists the answers in the quickfix window ranked by probability.
View on GitHub →Hands-on review
Splits a buffer into functions with treesitter, asks one question each, and ranks them into the quickfix list. Its accuracy table is deliberately blank.
Good for
Watch out for
Tested Sep 20, 2026 at a1577b0ec874 · Neovim in Docker; 68 tests via its own harness, run against a fake backend
How we reviewed this: we ran it in Docker with Neovim and its own test harness — 68 tests passing against a fake backend — and drove a real search over its fixtures. We made no Jev calls.
jev.nvim, by valentynkit, answers “which function does X?” inside the editor. Treesitter splits the buffer into functions, each becomes one question —
Does this function implement, or is it directly responsible for: “…”?
— and results land in the quickfix list ordered by probability, with anything below your threshold marked rather than discarded. Functions are packed into requests up to a token budget, four requests at a time, so a large file is a handful of calls rather than one per function.
Its README has an accuracy table with the numbers left blank, and a line saying they’re blank on purpose rather than filled with a guess. We’ve read a lot of READMEs this week; that’s the only one that does it.
Each question carries the function’s full source, truncated at about 3,450 characters, with no redaction. Nothing is written to disk.
A small editor tool that does one thing and says what it doesn’t know. If you live in Neovim and keep grepping for behaviour rather than names, it fits.
For the same idea over files, see semgrep and blink.
See how it compares with other tools in Best Jev tools, tested hands-on.
Review updated Sep 20, 2026. Numbers quoted from the project are its author's own; we don't publish our own measurements of Jev.
superagents-lab/jev-search
Search the web in plain language: Jev answers typed questions about your request, and the app uses those judgments to pick the query, the sources and the time range before ranking what comes back. Installable from the browser as an app.
crow-hq/llm-wiki-v2
An LLM-maintained wiki on Google's Open Knowledge Format with librarian and researcher agents, where a System One classifier (Jev via OpenRouter or TypeSafe, or a local server) routes each question to the right note. AGPL-3.0.
jexp/neo4jev
Navigates a Neo4j graph by having Jev score neighbouring relationships, then beam-searching for the most probable path.
New Jev releases, pricing changes, and the best new projects, once a week. No spam; unsubscribe anytime.
Powered by Buttondown. See our privacy policy.