Showcase

Built with Jev

Real projects running on Jev, with how each one uses it. Built something? Add it below.

Chrome extension

Slop Filter for X (PoC)

Scores every post and reply on X for AI slop and bot spam, then hides or labels it. Readers vote Slop or Human on each reply, which builds a labeled dataset to tune the filter.

How it uses Jev: One score question on a 5-level scale from "genuine human" to "obvious spam", plus a noul question that catches replies trying to talk to the filter ("ignore previous instructions…"). Verdicts are cached by tweet ID, so each reply costs one call however many people see it.

Not public yet

Apify actor

Jev Bulk Classifier

Label thousands of texts at once: support tickets, reviews, leads, scraped comments. Chain it after any scraper and every row comes back with typed columns and a confidence score.

How it uses Jev: Users define up to 10 choice, score or noul questions in the actor input. Each text is one Jev call with all the questions, run in parallel, and the answers become dataset columns you can export to CSV or Google Sheets.

Open on Apify →

Internal tool

Jev Tester

A small web app for trying prompts: write a state and questions, record the answer you expected, and compare runs side by side.

How it uses Jev: The fastest way to tune criteria wording before putting it in production. The playground on this site is based on it.

Try the playground →

From the community

Open-source projects other developers have built on Jev.

Chrome extension · open source

Slop Filter by Adam Roman

Hides AI-written posts and comments on X, LinkedIn and Reddit, and comments on YouTube. Watch it scan your feed live: each post turns green and stays, or red and folds away.

How it uses Jev: Asks Jev 19 narrow yes/no questions about how a post is written (the "it's not X, it's Y" pivot, three matching beats, a tidy bow on the end), then combines the probabilities with weights you can fit from your own labels. About 4 cents per 1,000 posts.

View on GitHub →

Chrome extension · open source

Slop Filter (Jev) by thonawit

Hides AI slop on X and LinkedIn and highlights posts about topics you care about. Strictness slider per platform, an allowlist of handles, and a diagnostics panel with the last 20 decisions.

How it uses Jev: Each post is judged once by Jev; the extension's own code decides what to hide from the probabilities. Answers are cached, so moving the strictness slider re-scores for free. About 200 ms per post.