Jev

Command Palette

Search for a command to run...

aitools.fyi picks · 488 listings · 131 deep dives

The sharpest map of Jev tools

Find SDKs, gateway hooks, MCP servers, agent routers, demos, and weird games built on TypeSafe's System One model. Real links, real repos, zero chatbot filler.

488

Listings

131

Deep dives

357

Card-only

10

Categories

139

With demo

408

GitHub repos

Jev spec sheet v1.13
Modeljev-latest
EndpointPOST https://api.typesafe.ai/v1/systemone
InputState plus typed questions, evaluated in parallel
OutputTyped answers, per-option probabilities, confidence
Latency70 to 500 ms end to end (per TypeSafe)
Pricing$0.042 / MTok input, output free

Primitives

choice

Pick one option from a list

choice, probabilities, confidence

score

Rate state on a rubric

score, probabilities, confidence

noul

Is this statement true?

noul (0 to 1)

Staff picks

See all featured
IntegrationsTypeScript5,253
DemoFeatured
Vercel's open agent framework, which ships Jev as the default evaluation model in its experimental evaluate path.
integrationsagentseval
Browser and computer usePython4,799
DemoFeatured
Browser Use's ultrafast agent: Jev picks the operation and DOM element in one request; a small LLM only writes text when typing is needed.

by @gregpr07

browser computer usebrowser-useagents
Agent toolingTypeScript2,735
Featured
Claude Code plugin that replaces the compaction summary with Jev decisions: every tool call and result is scored in one fast request, sta...
agent toolingcompactionagents
BenchmarksPython1,491
DemoFeatured
Semantic ifs from open models, on a 3090 at home. Independent; not affiliated with Jev or TypeSafe.

by @hhkkmon

benchmarks
BenchmarksPython851
Featured
Train a small model that chooses among a changing list of text options, one probability per option in a single pass. Includes Doom, chess...

by @hhkkmon

benchmarksgame
IntegrationsTypeScript796
DemoFeatured
Vercel Labs terminal CLI that can run Jev as the evaluation model for its evaluate command.
integrationseval

Collections

Shortcuts that actually save time

Browse by job

Categories

SDKs, MCP, browser agents, games, benchmarks, and the rest. Pick a lane, then filter hard on Explore.

Primer

Why Jev beats another LLM call

Jev is TypeSafe's System One decision model. You send structured state plus typed questions (Choice, Score, Noul). You get calibrated probabilities, not a motivational essay. Your code sets thresholds. Your app runs the side effects.

FAQ

Jev questions, answered fast

Short, quotable facts for humans and search crawlers. Go deeper on Learn.

What is Jev?
Jev is TypeSafe AI's System One decision model for software. You send structured state and typed questions (Choice, Score, or Noul). The API returns discrete answers with probabilities and confidence. Your code applies thresholds and runs side effects.
How is Jev different from an LLM?
LLMs generate open-ended text. Jev answers fixed questions with explicit probability mass over options. That makes routing, moderation, and safety gates easier than parsing YES or NO from chat output.
How do I try Jev on Vercel AI Gateway?
Use model id typesafe-ai/jev through Vercel AI Gateway with the AI SDK, including experimental_evaluate flows. See Vercel's gateway docs and the eve agent framework for TypeScript examples.
What is this directory?
Jev Directory at jev.aitools.fyi is an independent index of SDKs, integrations, Jev MCP servers, demos, and apps. The explore index lists the full public catalog; only stronger listings get on-site detail pages under paths like /sdks/ and /tools/. Published by aitools.fyi. Not affiliated with TypeSafe AI unless a listing is official.

Shipping with Jev?

Send the repo or demo. We actually click the links. Star counts stay honest or they stay blank.

Add your project
By Rishit