askgrep
Semantic grep that asks Jev a yes/no about every function for questions no regex can express, with cost reporting.
Overview
askgrep is grep for questions you cannot write as a pattern, such as "builds an SQL query by concatenating a value that came from the request" or "retries without backoff". It reads every function in a tree and asks Jev for a calibrated probability per function, listing hits with file, line and score along with chunk, cache, token and cost counts.
The point is coverage: a coding agent would otherwise either read the whole tree, which is expensive, or grep a few patterns and guess. The README prices a sweep of its own 819 functions (116,783 input tokens) at about $0.0049 with Jev versus $0.12 to $0.58 with Claude models, because output tokens are free on a classifier. A --dry-run flag prices a sweep before running it, and a demo folder reproduces the README's example hits.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | Does this function match the description (e.g. builds SQL by concatenating request data)? | yes · no |
How it works
- Split the tree into function chunks
- Ask Jev one yes/no per chunk, with caching
- Sort by probability
- Print hits with location, score and cost
Strengths
- Reads every function rather than sampling
- Very low cost per sweep, per the README
- Dry-run pricing
Caveats
- Needs a TypeSafe key
- Pricing comparisons are the author's
Getting started
export TYPESAFE_API_KEY=... then askgrep "<description>" <path>At a glance
- Type
- Project
- Status
- beta
- Runtime
- CLI
- Has code
- yes
Use cases and integrations
How curated lists describe it
Grep for the questions you cannot write as a pattern. Reads every function instead of sampling a few. Powered by Jev, TypeSafe AI's System One model.
hellogumbo/awesome-jev
Frequently asked questions
How is it different from grep?
It matches behaviour described in plain language, scoring each function with Jev.
How much does a sweep cost?
The README prices 819 functions at about $0.0049; use --dry-run to price yours.
Can agents use it?
Yes, it is designed so coding agents can search exhaustively instead of guessing.
Related projects
Deterministic code-graph context for coding agents on Postgres, AGE and pgvector, with optional Jev or local re-selection of symbols.
MCP serverbetaPythonClaude Code plugin that adds Jev as an independent judge for review findings, debugging hypotheses, design options and search hits.
PluginbetaRetrieval layer for Codex that finds evidence locally, optionally ranks sanitized excerpts with Jev, and returns bounded, cited evidence.
ProjectbetaRust CLI that classifies Git commit diffs with Jev into bug fixes, security fixes with CWEs, and change types.
CLIbetaRustDeepSeek Harness jev_search tool that runs ripgrep, scores every hit line with Jev, and returns only the top rows above a cut.
PluginbetaJavaScriptCLI that asks one yes/no question of every function in a codebase with Jev and ranks them by probability: grep whose pattern is a question.
CLIbetaPython