shipwithjev

Catalog / Tools & apps

0144GitHub

guesswork

Fish-style zsh autosuggestions, ranked by Jev instead of by prefix.

FindMalek/guessworkREADME ↗
Counts real <code>git clone</code>s of this repo — every first-time <code>install.sh</code> run does one internally, plus some people who clone without installing, so treat it as a close proxy, not a byte-exact install count.</sub>
</p>

```sh
curl -fsSL https://guesswork.findmalek.com/install.sh | bash
```

- **AI-ranked, not prefix-matched** — understands abbreviations and fuzzy
  matches, not just commands that start with what you typed
- **Fish-style inline suggestions** — grey text after your cursor, accept
  with <kbd>→</kbd>, works in emacs and vi keymaps
- **Four providers** — TypeSafe or Cloudflare Workers AI run the real Jev
  model; Anthropic or Groq work as fallbacks if you'd rather not sign up for
  either
- **Runs anywhere zsh does** — hooks `zle` directly, no raw escape codes, no
  terminal-specific code path
- **Cheap** — ~$0.000081 per request on Jev; a few hundred requests a day for
  a year lands around $10 total
- **No build step** — plain TypeScript, run directly by Node

**Jump to:** [Install](#install) · [Choose a provider](#choose-a-provider) ·
[Configuration](#configuration) · [How it works](#how-it-works) ·
[What it costs](#what-it-costs) · [CLI](#cli) · [Contributing](#contributing)

Prefix-matching history search only works if you remember how a command
*started*. You know you ran something with `--dir blog` in it last week, but
not whether it began with `amp`, `npm`, or `git`. guesswork closes that gap:
as you type, it asks an AI model which recent command you're most likely
retyping — abbreviations (`gst` → `git status`) and fuzzy matches, not just
literal prefixes — and shows the best guess inline, fish-style, grey text
after your cursor. Press <kbd>→</kbd> to accept it.

```
% gst▌  ⇢ git status  [0.990]                        fuzzy mode:

Also filed under Tools & apps

  1. 0616

    Jevstiller: distill Jev calls into a local model

    A self-hosted proxy that trains a local model on Jev's answers. It served 72% of Banking77 locally at 15 ms, with a 2% audit and a bound on disagreement.

    tomerglick57 · Tools & apps · 15 ms local

  2. 0611

    Scruple: code checks for what linters miss

    Semantic-aware linting with System One models: rules you can state in a sentence, checked against your code by Jev.

    @nalexpeae · Tools & apps

  3. 0606

    Algolia-style Postgres search with a Jev reranker

    Typo tolerance, typeahead, barcodes and facet counts in plain Postgres SQL, plus an optional Jev step that asks one question per top-10 result.

    @alex_forman1 · Tools & apps

  4. 0604

    Virtual Closet: outfits from closet videos

    built a virtual closet & outfit generator from raw closet videos. classification + semantic search thanks to jev 🥰. clueless closet dreams finally coming to life

    @maitreemervana · Tools & apps