Launch newsOriginal video
Jev launches: a System One model for decisions
TypeSafe founder Diogo Almeida introduces Jev, its RLCD training method, and its focus on software decisions. The post includes a 2:56 launch video.
Loading original X post…X
THE JEV GUIDE · CURATED EDITION
Jev Guide.
Launch news, explanations, demos, and community projects around TypeSafe Jev, organized by topic. Every entry links back to its original X source.
Curated selection · Not exhaustive · Updated manually845
curated posts
542
original videos
01 / EXPLORE
Explore the library
Showing 845 posts
0 means awaiting a verified post
How it worksOriginal video
Jev versus text-generating models
Jev does not generate free-form text. A side-by-side video compares parallel decision-making with token-by-token text generation.
Loading original X post…X
How it works
Pricing, evaluations, and the name Jev
The launch thread explains code-and-AI workflow evaluations, the input price announced at launch, and free output tokens.
Loading original X post…X
GamingOriginal video
Real-time Doom demo powered by Jev
A game demo shows real-time decisions. Diogo says it makes about ten Jev calls per second; the post includes a 2:43 video.
Loading original X post…X
Knowledge & searchOriginal video
Wikipedia link race: choosing among thousands of options
The model chooses the next step from hundreds or thousands of links, demonstrating fast decisions with constrained outputs. The post includes a 4:04 video.
Loading original X post…X
Developer resources
Jev technical blog and early access
The end of Diogo’s launch thread links to the official technical blog and early-access entry point for further reading.
Loading original X post…X
Launch news
TypeSafe announces access to Jev
TypeSafe’s official account ends its stealth phase, invites developers to request access, and quotes the founder’s launch post.
Loading original X post…X
Access updates
First Jev waitlist invitations go out
TypeSafe says it emailed a first batch of waitlisted users and will invite more people as access expands.
Loading original X post…X
Developer resources
Getting started with Jev: waitlist and Agent Skill
A community guide walks developers through requesting access and installing the TypeSafe Agent Skill.
Loading original X post…X
Community builds
Awesome Jev: early projects and use cases
A community-curated list gathers early Jev projects and ideas for developers who already have API access.
Loading original X post…X
Cybersecurity
Jev and CodeGraph for security review
The typesafe-security-review project combines Jev’s probability judgments with a code graph and OWASP/CWE data for low-cost security scanning.
Loading original X post…X
Access updatesOriginal video
Jev Waitlist Open: Reply to Get Access
Kevin Zhang invites Jev waitlist users to reply with what they want to build to get access.
Loading original X post…X
How it worksOriginal video
A hype-free explanation of Jev: a really smart switch statement
Nathan Flurry says Jev does not replace GPT or Claude; it is a really smart switch statement that can make many workloads insanely fast, cheap, and accurate.
Loading original X post…X
How it worksOriginal video
Jev Core Idea in 45 Seconds
The author shares a 45-second TL;DR video on Jev, finding the core idea simple but the original video hard to follow.
Loading original X post…X
FinanceOriginal video
Building an on-chain trading bot with Jev
A developer uses Jev to decide buy or sell and executes real trades on Kuru's on-chain order book via Monad every 300ms block. Link
Loading original X post…X
How it worksOriginal video
Jev can't generate text, but it has a lot to say
A user gives Jev a few hundred common English words and punctuation to choose from, showing Jev's selection-based expression beyond text generation.
Loading original X post…X
AgentsOriginal video
Browser Use + Jev: Ultrafast Browser Agent
Gregor Zunic showcases an open-source browser agent combining Browser Use and Jev, finding flights in 7 seconds for $0.0039, using a new action space per step, DOM state space, and small LLM fallback for typing.
Loading original X post…X
AgentsOriginal video
Custom Jev-style models can make agent systems more token efficient
Scott Williams predicts custom Jev-style models (parallel constrained decoding) will make existing company agent systems more token efficient, creating millionaires.
Loading original X post…X
AgentsOriginal video
Jev Benchmark: 500 Real-Time Agents Running in Parallel
Preliminary benchmark shows Jev running 500 real-time agents in parallel in a 3D environment, with 500ms average latency and 35 API calls/s using a naive implementation.
Loading original X post…X
Data analyticsOriginal video
Simulating Product Adoption Intent of 150 Fictional Personas via Jev API
The author built a system that sends 12 questions via the Jev API to each of 150 fictional personas to ask about their product adoption intent. The actual cost was 1.8 yen and it took about 5 seconds—fast enough, though latency is included.
Loading original X post…X
RoboticsOriginal video
Using TypeSafe Jev to make a Rubik's Cube solve itself like a human
A developer used TypeSafe Jev to make a Rubik's Cube solve itself with a beginner method in 94 moves, like a person rather than the 22-move optimal solution; Jev is not an LLM and answers one question with a probability in ~250ms.
Loading original X post…X
Workflow automationOriginal video
Jev Enables Real-Time Voice Control of Browser
A user interacts with Jev by voice; the transcript is sent to Jev, which returns probabilities in about 300ms, and the browser clicks accordingly, at roughly $0.0002 per decision.
Loading original X post…X
Developer toolsOriginal video
Reviewing PRs with Jev: one call returns 14 typed checks
A user shares using Jev to review real PRs: paste a diff, make one call to TypeSafe AI, and get 14 typed checks (hardcoded secret, SQL, etc.) as probabilities, claiming far lower cost than Claude.
Loading original X post…X
Developer toolsOriginal video
Unclutter: A Smart Ad + Slop Blocker with Jev
Unclutter is an open-source, free smart ad and slop blocker that uses Jev to automatically clean pages of ads, cookie banners, upsells, and annoying dialogs. BYOK.
Loading original X post…X
Workflow automationOriginal video
Jev Enables Computer Use Without Screenshots
Jev uses a local CoreML model to segment buttons and UI elements on screen, with on-device OCR reading labels, enabling computer use without screenshots or LLMs.
Loading original X post…X
GamingOriginal video
Jev Plays Subway Surfers at Superhuman Speed and Runs 50 Games at Once
Jev plays Subway Surfers at superhuman speed and runs 50 games at once, costing less than a cent per run. Jev does not replace LLMs like astra or fable, but opens up an entirely new world of capabilities.
Loading original X post…X
MarketingOriginal video
Jev analyzes 724 live ads in 40 seconds
Jev broke down 724 live ads from 37 brands in 40 seconds, covering hooks, formats, offers, CTAs, awareness stages, and landing page mismatches, using only 9 cents of tokens, coming to @stealads and MCP.
Loading original X post…X
AgentsOriginal video
How Jev Makes Agents 10x Faster and Cheaper
The post explains how Jev enables faster, cheaper, and more reliable agents through model routing and computer use.
Loading original X post…X
Data analyticsOriginal video
Analyzing 3,282 X posts with Jev to find growth patterns
Ian Nuttall used Jev to analyze 3,282 X posts (100M views), consuming 4,252,330 tokens for $0.1282 in an 8m 34s run, with 8 questions per post on topic, hook, tone, etc., finding how-to posts got 150 median likes.
Loading original X post…X
Launch newsOriginal video
jev(): PostgreSQL extension for natural language search
A PostgreSQL extension that searches your whole database in natural language via a single function jev(), with no index or embeddings, judging 129 rows in about 1 second.
Loading original X post…X
How it worksOriginal video
TypeSafe AI's Jev: A Different Type of Model That Doesn't Generate Text
Introduces TypeSafe AI's Jev model, involving RLHF co-creator Diogo Almeida, highlighting that it doesn't generate text but takes state inputs, making it a different type of model from ChatGPT.
Loading original X post…X
Developer toolsOriginal video
Instant Context Compaction with Jev
A user shares an ideal use case for Jev: instant compaction. By scoring every tool call and dropping irrelevant content, Jev can make compaction instant instead of relying on summarization prompts.
Loading original X post…X
How it worksOriginal video
Jev-style local LLM engine: judgment-only inference in 40ms
Taro L. Saito notes that, like Jev, a local LLM engine achieves 40ms inference by abandoning text/JSON generation and specializing in judgment, running on DGX Spark.
Loading original X post…X
Developer resourcesOriginal video
Full Jev Tutorial: Intro, API Setup and Three Demos
This tutorial explains what Jev is, how to build with it, and demos a voice-controlled browser, AI memory, and a YouTube predictor, including API setup.
Loading original X post…X
Developer toolsOriginal video
Jev Use Case: Scoring Tool Calls to Compact Sessions
Assign scores to all tool calls in ClaudeCode or Codex session logs and remove unnecessary ones to achieve immediate/compact functionality.
Loading original X post…X
MarketingOriginal video
Jev + SuperX: Predicting Virality with 61 Questions
Jev combined with SuperX generates 61 questions per post in ~1s for ~$0.0004. Trained on 9,481 real posts from 207 creators, it picks the viral post 2 in 3 times and never rewards reply bait. Free, no signup.
Loading original X post…X
GamingOriginal video
Jev Plays Smash Bros Against Itself, Controlling All 4 Characters
A demo shows Jev playing Smash Bros against itself, controlling all four characters and deciding the best move within a fraction of a second, using over 22 million tokens for the match.
Loading original X post…X
Developer tools
Jev Model as a Claude Plugin Compresses Context
A user shares using TypeSafe AI's Jev model as a Claude plugin to review unnecessary tool calls, compressing context from nearly 1M to 86K tokens in about 1 second.
Loading original X post…X
How it worksOriginal video
JEV Model Goes Viral: No Chat, Only Choices
A model called JEV that can't chat has gone viral, giving only three kinds of choice-based responses instead of writing code or explaining itself.
Loading original X post…X
Developer toolsOriginal video
JEV Co-founder Retweets: The Right Way to Use JEV in Claude Code
JEV co-founder @CompleteSkeptic retweeted and liked, calling it the right way to use JEV, noting the pain of system freezes when context fills up in Claude Code.
Loading original X post…X
GamingOriginal video
Jev vs Claude Haiku 4.5 in Tetris
The author pitted Jev against Claude Haiku 4.5 in Tetris with identical block spawn order and initial board, asking each to choose from up to 12 placements and dropping the block as soon as a response returns, to see who clears 20 lines first.
Loading original X post…X
Developer toolsOriginal video
Chrome Extension Using Jev to Skip YouTube Sponsor Segments in Real Time
A developer shares an open-source prototype Chrome extension that uses Jev to listen to YouTube audio, detect sponsor segments, and skip them in real time, costing about $0.005 per video with BYOK. Link
Loading original X post…X
Workflow automationOriginal video
Voice-Controlled Computer Use on Mac with Jev
A developer built voice-controlled computer use on Mac with TypeSafe's Jev, where dictated commands open apps almost before the sentence ends.
Loading original X post…X
Launch newsOriginal video
Jev's Unlocked Use Cases Are Underestimated
The author says he has been waiting for something like Jev since early ChatGPT models and predicts the fastest-growing SaaS by MRR in history within a month or two.
Loading original X post…X
Developer toolsOriginal video
Building a Faster 'Jev Use' Computer Operation with Codex+Jev
The author tried Codex+Jev to build an enhanced computer use called 'Jev Use', claiming it is faster and smoother than the built-in option with roughly comparable performance.
Loading original X post…X
MarketingOriginal video
JEV predicts performance of 700 outreach messages in 40 seconds for $0.09
A user shares using JEV on 700 high-intent leads and personalized outreach messages, predicting each message's performance, assigning confidence scores, and detecting lead-message mismatches in 40 seconds for $0.09, with lead scoring and buying analysis also possible.
Loading original X post…X
Access updatesOriginal video
Jev Explainer Video Surpasses 1M Views
The author shares that their Jev explainer video surpassed 1M views, was translated into many languages, and even got a Chinese remake.
Loading original X post…X
Workflow automationOriginal video
Building a macOS Downloads Folder Organizer with Jev
Marcel Pociot shares a macOS app built with Jev that monitors the Downloads folder and applies customizable rules to rename and file downloads, such as detecting invoices and moving them to a dedicated folder, with no other LLM calls involved.
Loading original X post…X
Access updatesOriginal video
How to Use Jev: 10-Minute Setup Guide
Steps to get started with Jev: join the waitlist and install the official skill so your agent writes correct calls. Link
Loading original X post…X
Launch newsOriginal video
Jev launches Website to App: turn any website into a native mobile app
Just paste a URL: jev-1.13.0 decides how to build the original website as a native mobile app, then shipper submits it to the app stores for you.
Loading original X post…X