View demoFeed cleanup
Classify elements, clean up a page
A browser-extension experiment identifies advertising elements in the DOM and hides the elements it flags.
Everyday interactions
Make information easier to act on: classify incoming messages, evaluate candidates, retrieve relevant records, or suggest the next edit. Each experiment connects a user’s intent to a structured choice.
Explore by application
7 curated examples
02
Everyday interactions
Filter, organize, and retrieve the information that matters.
View demo
View demoClassification
Typed JSON on a local model
A related Qwen-based experiment predicts categorical outputs for multiple JSON keys together on an Apple Silicon laptop.
View demoSemantic search
Query data in everyday language
A PostgreSQL extension applies natural-language criteria to rows, exploring semantic filtering inside the database.
View demoClassification
AI Email Classification for Inbox Organization
A community demo uses Jev to classify a batch of emails, exploring structured decisions for inbox organization.
View demoRanking & evaluation
AI Pull Request Risk Review
A PR-review experiment evaluates a diff against typed checks and maps the results to a review outcome, escalating ambiguous cases.
View demoInput assistance
AI Next Edit Suggestions for Code
A community prototype explores a Next Edit Suggestion feature using Jev to anticipate a developer’s next change.
View demoSemantic search
Semantic Log Search for Incident Investigation
A semantic log-filtering experiment exposes a CLI for agents and a terminal interface for people to find relevant events.