Updated Nov 10, 2026
855 projects, posts and guides about Jev, the System One model from TypeSafe AI. Each entry links to its source and shows the cost and speed its author reported.
Diogo Almeida
@CompleteSkeptic
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI? I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev • 20-200x faster • 40-400x cheaper (w/ output tokens free) • Frontier composable intelligence optimized for decisions AFAICT the shortest path to AI-based economic revolution
Newest first, 864–896 of 1072. All 1072 in curated order
Guidenishtahir.com
Nish Tahir builds one from Qwen3-1.7B: constrain the output to a fixed set of options, read the probabilities in one pass, then fix the overconfidence with temperature scaling. It comes with a repository of scripts to build a dataset, evaluate, fine-tune and calibrate.
Jason Lu
@jasonlu_ai
JEV is changes the world of E2E testing! Same eBay test flow, completed-run medians: Jev: 47s / $0.0067 GPT-5.6 Luna: 62s / $0.0277 Claude Sonnet 5: 79s / $0.4062 Try jev-e2e. github.com/perixtar/jev-e…
Chandramouly Kandachar
@chandamamz
@typesafeai's Jev controls the 2 hands and each finger to play the piano in real-time. Jev only "sees" what we see and plays this from the "note waterfall". It uses @browser_use's jev-ultrafast and some decision scheduling to make this happen in real-time. Sound on 🔈🔉🔊
XGames and real time
Ateeq
@TPateeq
I used Jev to solve a problem every agent eventually runs into: reading logs. 22.8M lines, and running an LLM on every one would've cost $1,120. Tocsin groups them into 11,812 repeating patterns, then asks Jev about each pattern once. 6 minutes, 64 cents, 123 patterns that actually needed to be looked at. http://github.com/TPAteeq/tocsin The paging policy is just a prompt. You tell it what should wake someone up at 3 am and what's just another log line.
Guidejahanzaib.ai
Jahanzaib’s guide to moving a Jev integration to Microsoft Decision-1. The endpoint path ends in /v1/systemone, the request has the same state and typed questions, and the price is the same $0.042 per million input tokens with free output.
Everton Carneiro
@everton_dev
I built a tool that finds App Store keywords by reading the competition, and uses Jev to judge them. What it actually does: 1. Turns the app's own listing into a handful of search queries, with Jev filtering out the ones nobody would type. 2. Runs those searches on the App Store. Whatever ranks is the candidate pool. 3. Jev judges each candidate: is this really an alternative to the app, or does it just share a word? The lookalikes get dropped. 4. The strongest survivors become the competitor set it mines for keywords. No competitor list to maintain, nothing hallucinated: competitors are whoever Apple already ranks, minus the ones Jev rules out.
XSEO and GEO
Aditya Singh
@xyz04274951
I made this for fun. Wired a mic over Fusion 360: click, speak, Fusion runs the feature. @typesafeai ‘s Jev only decides if the utterance is a command.
XRobotics and devices
Kun Chen
@kunchenguid
almost every day i hear people ask "when should i /compact my session" there's no easy answer because it depends on how likely your future action will need detailed context in the existing window but we have Jev now! introducing compact-adviser - an agent plugin you can use in claude and pi today to help determine whether you're likely at a task boundary that's safe to compact https://github.com/kunchenguid/compact-adviser i built a private eval set from 40 real sessions and manually labeled all the safe vs unsafe checkpoints to evaluate this, and hillclimbed the Jev prompt till it performed quite well i also made it so that the classifier will - optimize for precision (not triggering a compaction prematurely) when context window is small - and gradually shift to optimize for recall (not missing an opportunity to compact) when context window fills up, because the cost of not compacting becomes higher, and at the end the agent will be forced to compact anyway it supports a "hint" mode (just give you a hint and it's up to you to run /compact) vs "auto" mode which runs compaction whenever Jev says it's safe to do so if you have Jev and want to put your compaction on autopilot, try this out and let me know how it goes! support for more harness is coming soon as well
GitHubBenchmarks and evals
Turns human labels into calibrated Jev classifiers with GEPA.
Sutro
308
GitHubOpen source
A 151M non-autoregressive decision engine, with its numbers published.
Heman10x
294151M
GitHubBenchmarks and evals
Nine experiments and 28 predictions, all fixed before any data.
Will Kelly
0123,805
GitHubOpen source
Typed decisions over images, not text descriptions of images.
Andrue Anderson
3
GitHubSecurity and abuse
Bug-hunting crawlers built from five Unix-style primitives.
russfranky
GitHubCoding and code review
OpenAPI compatibility checks, structural and semantic.
Shuhan Sun
5
GitHubAgents and browsers
Voice control for macOS where Jev is the fallback, not the router.
Coco Research
2
GitHubRouting and model choice
A Claude Code and Codex skill that picks the model for a prompt.
heyman333
1
Jon Kraayenbrink
@kraayenJon
jev is INSANE. in 243 ms it checked a website for 35 tells of ai slop. purple gradients. emoji headers. "seamlessly". fake testimonials. bento grids. the works. used $0.00015 of tokens. paste any url, get a slop score. free: madewithjev.com/free-tools/ai-…
Hamilton Ulmer
@hamiltonulmer
I made a DuckDB extension where you can use @typesafeai 's Jev to do quick classification of rows in any csv/parquet file or duckdb table about 10sec for 1k rows ~ better than using an LLM, way more ergonomic than a classifier game-changing for data analysis!
Steve Krouse
@stevekrouse
typesafe's jev is fun! live demo you can play with: typesafe-demo.val.run
GitHubAgents and browsers
A mobile agent on a real Android phone, with Jev making each decision.
Droidrun
4479 actions in ~21 s
SkillAgents and browsers
A Codex skill: Jev clicks, Codex thinks and checks the result.
coliney
1.2k~5–10×
GitHubSales and leads
A LinkedIn recruiting agent that browses profiles and saves the evidence.
skeptrune
52
GitHubSecurity and abuse
A judgment layer that checks every Claude Code tool call against your policy.
Clownware
1~$0.04/day
GitHubAgents and browsers
A gate on every agent tool call, with Jev as one of its sensors.
somoore
0
GitHubGames and real time
Pong where the ball moves one step per model decision. Jev against the LLMs.
Ably Labs
1
GitHubGames and real time
Jev plays real-time Tetris against Claude Haiku or Gemini Flash.
Tony Dinh
20
GitHubGames and real time
A solo mahjong coach: Jev picks the tile to discard. In Japanese.
smilior
0
Same directory, different order: the entries added most recently, rather than the curated order the home page opens in. The date on a card is the day it was added here, not the day its author published it — this directory is newer than most of the work in it.