Tools & apps

We got access! And in this video we explore @typesafeai's Jev, our first hands-on, then build Excuse Court with Codex: structured questions in, probability estimates out. Cat trouble or alien abduction: what would you test with it? https://t.co/32iOhoIXFC

I made a tool that allows customers to create their own filters 👀 with JEV this is the first time it actually feels instant > type "for the winter" and it classifies every product on the fly, based on the product information with EasyFinder we've been chasing intent-based search for 18 months, so this is actually really cool! still an experiment, but having a few other ideas on JEV for i will shar

Got early access to @typesafeai Jev and spent a few hours building on it. The problem I keep hitting: agent traces tell you which tool ran, how long it took, what it returned but never whether the agent is actually getting anywhere. An agent editing, testing and reverting the same file six times looks perfectly healthy in the logs. So I made JevScope. It watches an AI agent work and asks Jev what

i was curious if @typesafeai 's jev would make different judgements if you push a problem through a Choice vs. a Noul, so i ran it through a short series of trolley problems the only time they disagreed on judgement? deciding whether or not its worth killing more people to score higher on a benchmark 😬

built a chrome extension that lets me write my own X algorithm. it hides engagement bait, empty hooks and ads before I see them. the rules are plain English. Jev by @typesafeai checks every post in half a second. anyone want to try it? https://t.co/ybO4pSm1pK

Jev 场景探索：基于高质量数据集做个性化推荐。 孩子或家长输入任意的自然语言，基于整个数据集来给出推荐结果。 初试下来，命中的结果高度贴合预期，平均耗费时间较长，需要 15 秒左右，故展示效果的等待时长经过删减。 https://t.co/gNLi5Qs9Mo

Just trying out Jev, I made a FlightWifi Chrome extension that: - Checks Wi-Fi on your flight (automatically) - Detects if the aircraft supports video calls, email, or nothing at all - Detects Starlink or traditional Wi-Fi and shows the verdict directly on Google Flights and more - So you know if you can actually work before booking ✈️ No more guessing if the Wi-Fi is worth paying for https://t.co

Small demo built with @typesafeai Jev and @cactuscompute Needle Coding by @claudeai Fable, maps by @openstreetmap, additional information by @mapillary, voices by @Alibaba_Qwen Qwen3 TTS, and music - which I did not ask for, btw - by @StepFun_ai ACE-Step 1.5 Three prompts total. "dude, I have no idea what we are going to do here... but 1) not a kortexa repo yet, so no factory, pure vibing, 2) grab

My god.. so many nice use cases for the new Jev model. I quickly built a PII and sensitive data redactor. First checks each sentence if it needs redaction, then feeds the affected sentences individually to pick the words that should get redacted. Compared to e.g. using regex this catches also information that is not following a known structure but should still be redacted (like a business secret o

I just dropped JevQL, semantic WHERE clauses for vanilla Postgres, powered by @typesafeai Jev. No extension and nothing to ask your DBA for https://t.co/vJgUBnguXi

Yes yes, Jev can drive, Jev can be a programming language... but lets just skip the middle man ... Jev be the "computer" !!! Motorola 6502 emulation directly on Jev 🤣😅 @typesafeai @dotpem Took a bit "hill-climbing to figure out how to instruct and the data shape, but now it runs whole (very) short programs. As you can see it still gets it wrong against the programmatic emulator reference every onc

Used Jev to build voice builder in Youform. It is so fast! https://t.co/hPFKW3wOOL

Concrete real world use case with JEV from @typesafeai Filling PDF forms: ~1.5s to fill, ~500ms to classify - insane numbers Especially useful when the input data is dirty (bad fields, wrong fields), with a human in the loop to review and fix https://t.co/kDUure64IG

Jev + Grok + 青空文庫のデータを使って、新たな青空文庫風の小説を無限に生成するアプリを作ってみた 青空文庫のテキスト断片をランダムに抽出して Jev が続きの文章として使えるかどうかを判定、OK なものを Grok がアレンジして繋げていく形 https://t.co/Dqo1YE2a5M

Jev playing the piano. I gave it the key, all 88 notes to choose from and the sequence so far. Did chords in a second pass once we had the melody by giving it the bar's notes & 7 chords to choose from. Plays it safe & loves alternating between G and A but with some lessons, it might get there.

I built a linter for AI writing tells, and the judge is a model that can't write a sentence. Sniff Test reads a draft one paragraph at a time and asks Jev ten yes-or-no questions. Is the claim hedged three times. Does the closer just restate the paragraph. Is there a not-X-but-Y turn. Is there a cost figure with no price next to it. It comes back with one probability per rule in about a fifth of a

I wanted an easier way to test my jev requests and see if my results are improving, so I built jevals! It's a local workbench for testng the Noul, Choice, and Score primitives. You can mix requests as well. You can run a seed command to see and play with some of the examples in the @typesafeai Made a quick demo video if you're interested in testing jev in a million different ways lol

I wanted a feed I could describe in plain English, so I hacked one together. Two prompts control my YouTube and X feeds as I scroll. Jev responds in 70–500ms. Estimate: $0.50–$1/month for 8h/day. A fast, cheap decision primitive alongside LLMs. https://t.co/oD3koQ4Nab

one more for the pile, from the writing side: Sniff Test, a prose linter where Jev is the judge. ten yes/no questions per paragraph, 182 ms median, and on 54 clean paragraphs it raised 1 false flag where Haiku raised 37. runs as a commit hook or a Claude Code skill. https://t.co/fkN2VDyoSJ

Here's what happens if you let Jev draw for multiple iterations. Every pixel gets the color of the neighbors from the last run and a scaled down map of the whole image as the context. Worked not that bad until the 6th or 7th generation when it started to destroy the sun again 😂 https://t.co/RhM5y7laNn

Inspired by @typesafeai's Jev, I built sarvam-jev - an open engine that reads typed decisions straight out of an Indic model's logits instead of generating JSON token by token. Same sarvam-1 weights, 8.8× faster than making it write the same answers — and the output can't be malformed, because nothing is generated. Runs entirely in your browser. No backend, no waitlist: https://t.co/jLnMT8EeDq Hin

Jev lets me search my physical library by looking at the entire book index in parallel in 300ms. Here, the answer to “Who invented photography?” is probably on page 98 under "Niépce, Joseph Nicéphore" https://t.co/15HwgDZ67D

Put Jev in my TV app (stealth for now). Netflix has 2,000 engineers and still can't do "something light, no gore, we're tired." The trick isn't the model. It's refusing to let it generate anything. It ranks what I already have. Pick a mood on the remote → it re-ranks 797 movies & shows in one pass. Cards glide to their new rank. 104,557 tokens · $0.0044 per press · ~1s server-side, 1.5s on a cheap

誕生日プレゼント５０個の仕分けにJevで API費用は0.7円くらい。 https://t.co/we9PmlcdYC

今読んでる文章の段落に応じて、Jevで動的にBGMを自動生成するのを作ってみた。コード進行をいい感じにする。 https://t.co/tr3xR9PzXe

I built a linter for AI writing tells, and the judge is a model that can't write a sentence. Sniff Test reads a draft one paragraph at a time and asks Jev ten yes-or-no questions. Is the claim hedged three times. Does the closer just restate the paragraph. Is there a not-X-but-Y turn. Is there a cost figure with no price next to it. It comes back with one probability per rule in about a fifth of a

been playing with Jev from @typesafeai and built unsaidbrief the idea: we give AI builders vague prompts, then spend hours fixing decisions we never actually made paste your brief and it uses Jev to flag missing decisions and possible conflicts, with probabilities you can see answer what matters, recheck, then take the clearer brief into your builder small walkthrough below. curious what it catche

iPhoneで動くVLMベースのjev-likeを作ってみました。学習なしです。 https://t.co/wiaGAQ1Sim

Drive-thru ordering at Big Kahuna burger powered by @typesafeai Jev The model is really good/fast/cheap at Natural Language -> Menu -> Selection https://t.co/Q2b0DRH9pV

went deeper down the Jev rabbit hole 🌻 196 artworks from The Met mapped between quiet ↔ loud and sacred ↔ domestic change the words and watch the collection rearrange https://t.co/deUdeapWKu

Can Jev escape a maze? Turns out it depends how you write the map. When given as a 2D grid, “wall above?” is worse than chance. “Wall to the right?” is 99.9%. Rotate the page and every side becomes 1.00. It isn’t seeing the grid. It’s reading the next character. So I gave @typesafeai’s Jev the maze as lines of sight and named places, not as a map. Notes below.

love jev, but upset it - isn't open source? - it lack vision capability? We fixed all of that, introducing https://t.co/DlNeKo2k5e A fully open source library which takes any HF model and Jev-ify it, with an API endpoint Now on github, and live in production at @FeatherlessAI https://t.co/dFKBbR411y

ローカルJevできた。確信度ないけどｗ 40msで返事返ってくると不安になるけど一応GPUはピクっと動いているのでLLMの出力にはなっている https://t.co/8eEtZ0CpSU

I got tired of choosing between X's firehose and someone else's idea of relevance, so I built Your Signal. It uses @typesafeai's Jev to score posts already on screen, then applies my rules locally. Reversible, BYOK, no telemetry, open source. https://t.co/1dNRESvT5J https://t.co/sYOVB3eIoy

**jev scroll break** took a stab at making a design skill inspired by @anshuc dream loop skill, and made taste loop! tested it on my portfolio cuz it needed an update and holy it looks much better now https://t.co/6z0lsmyHt0 https://t.co/lgspHpuKrA

Also been playing with @typesafeai Jev. Insane. A pile of apps suddenly become possible. What a time to be a builder. Sharing experiments here. First one: Keystroke oracle / predictive launcher A normal launcher ranks by aliases, fuzzy match, and habit. Jev reads intent. Type “the pdf I just downloaded” and the newest PDF is already the top hit. Full confidence on every keystroke. About 100 ms.

another demo for jev for ~instant variable generation https://t.co/ddjvvHXpai

Jev added more precise camera controls and ~230x faster planning to an editing tool I built on top of @mintdotgg! Jev converts natural language to precise camera movements (ie spelling ASTRA) by answering a few calibrated multiple-choice questions (move kind, plus direction and magnitude band per camera axis) By creating keyframes it emits a structured trajectory (time/azimuth/elevation/distance)

yes / no jev demo https://t.co/cINZEnqWFm

made some time to test jev from @typesafeai used theMealDB and sent each recipe to Jev to score on five dimensions: quick, cheap, healthy, impressive, and beginner-friendly, while filtering out desserts, sauces, bread, etc. The whole run took just over a minute and cost $0.0455. I ran it twice and got slightly different rankings. Close results swapped places over the runs. So you either need to us