Playground
CommunityJev Cookbooks
TypeSafe Typewriter
A live Val Town demo that re-asks 16 typed judgments about your text in a single API call on every keystroke, covering traits like passive-aggressive, AI-written, tone, and urgency.
- Format
- Playground
- Reading time
- 3 min
- Published by
- Community
- Last verified
- Sep 2026
Works forJev for Scoring
Used inJev for Software & DevTools
Highlights
- 16 typed judgments are asked in a single call and re-run on every keystroke.
- Rows include passive-aggressive, AI-written, tone, and urgency, and flash as values change.
- Runs as a hosted Val Town app using the Twind CDN, with no install step.
Watch out
A minimal live demo showing only the 16 named judgments, with no published calibration methodology for them.
Questions about this entry
Short answers drawn from the linked project or page.
The entry describes a hosted Val Town app that re-asks 16 typed judgments about your text in a single call on every keystroke, with rows flashing as values change. The captured page is essentially empty, just a title, so the source does not document the request shape, the model version or what per-keystroke billing adds up to.
Named examples in the entry are passive-aggressive, AI-written, tone and urgency; the rest of the list is not documented in the source. Since no calibration method is published for any of the rows, treat each number as something the model produced rather than a measurement, and expect the named set to change.
A demo, and that is its value. It runs as a hosted Val Town app using the Twind CDN with no install step, so you can see the shape of many typed judgments returned from one request. Nothing about persistence, batching, throttling or calibration is documented, so copy the pattern rather than the app.
Not documented, and it is the number to want before copying the pattern. The source does not say whether the app debounces, how long a keystroke burst runs, or which model version answers, so the price is unknowable from the page. If you replicate it, debounce, batch and cache per document.
The source says nothing about storage, logging, accounts or retention. The captured page is essentially a title, and the entry documents only the judgments and how they update, so treat it as a public demo: do not paste anything sensitive, and do not assume your text is discarded server-side.
Not documented. The page is a hosted Val Town app, and the entry does not say whether it runs on the author's key, a shared demo key or a key you supply. Since it fires a call on every keystroke, that is the first thing to confirm before typing anything long into it.
Not documented. The captured page is a title, and the entry says nothing about error states, retries, loading indicators or what a failed keystroke leaves on screen. Since a request fires on every keystroke, expect to test that behavior yourself before relying on the display.
From the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.