Jev.aitools.fyi

Command Palette

Search for a command to run...

DemoFeatured
Product profile
TypeScript5,253 starsUpdated 2026-09-17

eve

Vercel's agent framework with Jev wired into the experimental evaluate path, so scoring stays typed.

Overview

eve is Vercel's open, filesystem-first agent framework (agent/instructions.md, tools/, skills/, channels/). This directory indexes it because public eve materials describe Jev on the experimental evaluate path for typed scoring; confirm shapes on eve.dev for your pinned version.

Problem: Durable agent frameworks need evaluation hooks that stay testable instead of free-form judge prompts.

Built for: TypeScript teams building filesystem-first agents on Vercel who want structured evaluation alongside tools and skills.

First indexed on Jev Directory: 2026-09-17

Creator and team

Name
Vercel
Organization
Vercel

How Jev is used

Role in the product flow
Agent evaluation (experimental evaluate path per listing and eve marketing)
Primitives
ChoiceScoreNoul
State in
Structured evaluation payloads as documented on eve.dev for the evaluate integration (verify in your version pin).
Decision out
Thresholdable evaluation scores used in agent workflows instead of unparsed judge prose.
  1. Author agent under agent/ layout
  2. Call evaluate path with structured questions
  3. Apply thresholds in TypeScript policy code

eve keeps orchestration in conventional files while evaluation can call System One through the experimental path described in this site's catalog listing and on eve.dev. This profile does not duplicate eve's full evaluate API: read Vercel's docs for the exact question schema, models, and gateway configuration in the release you deploy.

Features and stack

Features

  • npx eve@latest init scaffolding
  • Filesystem-first agent layout
  • Channels and schedules for durable agents
  • Hosted docs at eve.dev

Stack

  • TypeScript
  • Vercel AI Gateway compatible flows
  • TypeSafe System One evaluate path

FAQ

Where is Jev configured in eve?
Follow eve.dev/docs for evaluate integration in your release. This directory listing summarizes Jev as the default evaluation model on the experimental evaluate path.

Related learn guides

Original Jev guidance that pairs with this product pattern.

  • Vercel AI Gateway Jev

    Try typesafe-ai/jev through Vercel's gateway and the AI SDK evaluate path.

  • System One model

    The model family behind Jev: parallel typed questions, one forward pass, probabilities you can threshold.

Related products

Hand-picked neighbors with rich profiles or overlapping tags.

Integrations
Demo
Workday Sana is the new Workday Home, powered by Jev, with real-time forms, calendars, approval cards, and charts drawn from the design system.
integrationsenterprisegenerative-ui
IntegrationsTypeScript
DemoFeatured
TanStack AI is the TanStack team's AI SDK surface.
integrationssdktypescript
Browser and computer usePython4,799
DemoFeatured
Jev picks the browser op and DOM target in one shot; a tiny LLM only shows up when something needs typing.

by @gregpr07

browser computer usebrowser-useagents
By Rishit