X posts

Okay so we also built something else. So we are using @Muse & @typesafeai Jev together now. I’m using Jev to grade my AI assistant Sarah on her own work before I ever see it. So (Jev by Typesafe) will act as the judge. Every caption draft gets scored against my actual writing voice: my hooks, my outro style, my banned words, all of it. And It's bigger than captions. Before she states a fact, the judge checks it against her own records, it caught a wrong number tonight before I saw it. Before she acts on a vague instruction, the judge decides: act, or ask me first. A checklist that never

using Jev to score 2,952 museum objects on "is this architecture?" https://t.co/GaJgJtMvdb

I found the "Internet" moment in using Jev Jev analyzed every LLM and chose the BEST-VALUE one for each prompt It cut my costs and time by ~80% - and it’s probably the best way to use Jev... here's how to setup it in 10 min: step 1 → open @typesafeai , create API key (keep it off chat paste) step 2 → store TYPESAFE_API_KEY in a local .env step 3 → clone the Github below, then npm install && npm test && npm run demo step 4 → test Jev before connecting an agent: npx tsx sdk-runner/cli.ts --route-only "Rename a button label" step 5 → dry-run first: npx tsx sdk-runner/cli.ts --dry-run "you

i've built an extension for twitter that ranks every tweet by relevance based on your interests using @typesafeai 's Jev, highlights very recent tweets to maximize engagement and assesses your tweets based on your goal (also using Jev) should I open source it? https://t.co/GgQjTT2NWy

I wanted to post a preview of our Jev benchmark dropping tomorrow. We are using Jev to generate symbolic programs (state machines) through an iterative process. The results so far have been amazing! The old process took ~2–10 seconds. With Jev, it’s ~0.3–2 seconds. The costs have also dropped by an order of magnitude. Jev is also used in control flow, removing the need for brittle heuristics.

I built a token compression engine using Jev. It evaluates document (huge text) chunks in parallel, in sub-100ms passes, stripping 85% of boilerplate so we only pay Claude or GPT-4 for high-signal answers. Here is a demo https://t.co/3nVjtG8zZY

noticed below cumulative impact of using Jev + Jevcache + Fable: - Per-decision: ~220–780ms Jev vs 20–40s claude before - Loop's decisions are ~50× cheaper; Fable was consulted only on the 3 low-confidence steps - jevcache: earlier identical-phrasing re-run showed cached (0ms replay) live.. repeat run was 26% faster (7.2→5.3 min)

Jev can outperform the rules many companies use to prevent fraud. We found that without any pre-training, it accurately detected 93% of a fraud ring. In contrast, an LLM in a similar set up only achieved 62%. https://t.co/ronCD7hLvb

3 days. First AI product I've ever built, with Jev doing some of the heavy lifting inside it. Here's the landing page. No Figma at all, just Cursor ai and the design skill I built. Launching soon https://t.co/VBhWBCVEMf

I just classified a month of my Gmail for eight hundredths of a cent. 150 emails → 9 folders. Read-only, nothing written back to my inbox. Built with Jev. Repo below 👇 https://t.co/PaUcxpcDE1

I builded Jev Reflex which can control your mac with hand gestures and voice. It is build with Jev 1.13 from @typesafeai . I builded it with @OpenAI Codex using Astra and Sol. It is using Jev credentials from @OpenRouter and for all this just used $0.01. As you can see in the video, it recognizes hand gestures and voice and can do various task on mack, like mazimize or minimize anything you point you index finger. And then pinch gesture to complete it. It took me 2 hours to build it on a monday morning, which included code, enabling controls on mac and calibrating it first. Project availbl

Using JEV by @typesafeai to classify the 100s of inbound deals we get, side-by-side with a small open model we’ve been using for our in-house AI rating module that self-trains. impressive how close JEV gets while being significantly cheaper and faster to run for classification :) https://t.co/6c6I9SHQuY

There's a fun mix of excitement and cope on my timeline re-Jev: - OMG I made Jev a year ago this is dumb - You can fine tune this encoder into a better, faster, local classifier, etc Re the "just train your own small classifier" crew - Voyage trains great, SOTA pointwise rerankers that are obviously specialized to the reranking task and have been at it for years, likely also using a strong synthetic data recipe... Exact setup in the response tweet. Do note that using Jev is definitely more expensive than using Voyage rerank-3 bc of the tokens in the lengthy rubric included in every request (t

Use @typesafeai's Jev model with Pinecone to rerank results with natural language criteria! Usually with rerankers, it's hard to cleanly specify what should and shouldn't be returned in results. Jev resolves this by refactoring the problem into evaluating against distinct binary criteria, which pairs great with Pinecone retrieval! In this demo, we compare using Jev and Claude to rerank 200 returned candidates from Pinecone. Jev returns a reranked list in about a second — 830 to 1,300 ms across eight test queries. Claude Opus 5, doing the same job in one long-context call, takes 4.2 to 6.8

First real use case with Jev got up and running last night. Used it with GLiNER 2.5 base to help dramatically speed up entity / relationship extraction. GLiNER by itself for Graph RAG over a huge document corpus is fairly unreliable. Using Jev to help determine what it got right and then send everything else to my local 35B model. ~25x performance boost.

simple real-time bug checker made with jev in case you code by hand sometimes also works with english (underlines inaccurate sentences) https://t.co/hJt7xzRLpB https://t.co/XsitOtjnnh

We just cracked how to make a great Meta ad using Jev ( a classifier model from TypeSafe AI. ) We pulled around 2,000 real ads from 8 beauty brands off Meta’s Ad Library and graded 302 of them with Jev across 9 dimensions: hook, format, funnel stage, CTA, claim risk and boldness. Jev is an AI evaluation tool from TypeSafe AI that lets you run the same structured review across hundreds of inputs. The best ad in the entire set was a drag performer climbing 300 stairs to see if her makeup survives, for e.l.f. Cosmetics. She puts the makeup to the test by climbing the stairs, and that’s the ad.

Built a Claude Code mod using Jev that watches what you're asking and routes the turn to the right model. Quick renames go to Haiku. Real implementation? Opus. Architecture planning? Fable. @claudeai @ClaudeDevs @bcherny @lydiahallie @trq212 https://t.co/DgIgIGpHLq

Built a Ticket Triage Copilot using Jev (@typesafeai) instead of a general LLM for support ticket classification. 🔗 GitHub: https://t.co/zQ9oB7t7Eq

This weekend I've built a tool using Jev, was a really fun thing to learn! It classifies your followers in seconds, so you can analyze your audience, check comments⬇️ https://t.co/4YmKUVVlBq

Your browser has 100 tabs. You're using 6. Meet Tab Bouncer 🚪 a web extension I built that checks every tab at the door and shows the freeloaders out. Built with Jev by @typesafeai: ⚡ 100 tabs judged in one call, ~1s 💸 a hundredth of a cent per sweep https://t.co/H4eMmMTM4S https://t.co/st0zISjA81

No, 'caliente, caliente no más! We shipped JarvisCore 1.12 and it comes with Jev by @typesafeai ! Come and learn about Jev next week on Wednesday. We're bringing @askmuyukani founder at @prescottdata for a live discussion on how JarvisCore is using Jev to build faster and cheaper agents, and what opportunities exist to extend Jev in agent harnesses.

I built Jev Dark Pattern X-Ray — a small experiment using Jev from @typesafeai to detect manipulaive UI patterns on mainly ecommerce sites. It also have a demo site to check how it works, but you can paste a real url and it will check for fake urgency, scarcity pressure and other tactics used my ecommerce sites. Code: https://t.co/5nqnd8TFrb

I was also looking into how we can integrate Jev into a real hardware and I built a custom firmware for my keyboard that predicts and highljghts the next key using Jev. Mostly real time. This is not a simple dummy 2kb model that lives in the board. In my project Jev gets context of the current active app on a screen so it can suggest actual vim commands completion! For real. And a real word endings or even phrases. lmao this experiments are getting out of hand this quite cool

Demo of using Jev to analyze spoken text in real time. Try it out at https://t.co/R4t6r110wI Tested in Chrome and Safari on Mac and iOS. Source code is at https://t.co/F5UaD8AcDz Thanks @typesafeai @CompleteSkeptic @notkevinzhang https://t.co/Hfa0JrWQMJ

Introducing Cursor for writing. Who says Jev can't generate text? Is Jev fast enough to finally solve auto-complete? Code is a lot easier to predict, so how does it work? It keeps a history of your previous 1000 sentences to predict the next word faster and better as you keep typing with binary trees to speed it up. As you type more, it gets better at predicting your next word until it actually accelerates writing. Built with Jev from @typesafeai

Holy shit. Using Jev in my JARVIS just made it like actual super intelligence. what the fuck. Will release the update asap. https://t.co/LJZA4z7fSf

I think Jev really flourishes inside your code more than in the actual product. Right now, I’m using Jev to classify the thinking mode of our AI agent, so it can select a high, medium, or low mode depending on the prompt.

I'm building a real voice-controlled remote for Mac. And I'm using JEV for almost everything. 🧠 Decision making — auto allow / deny actions 🖥️ Computer Use — interact with the Mac 🌐 Browser Use — navigate and control websites, much faster 🧩 Generative UI — dynamically generate forms and controls 🎙️ Voice Control — control the entire workflow remotely The problem is macOS TCC. When an agent triggers certain macOS permissions, you still need a real physical interaction. A Computer Use Agent can't simply click the approval button. That's intentional. Great security. Terrible for unatt

Humble beginnings using Jev. Trying to integrate it into Codex and help with speeding up browser use. Been working on it all day and running hundreds of trails. So far I've spent... $0.00725 https://t.co/yX3mPB41pL

Built for my mum, who has MND / ALS , using Jev @CompleteSkeptic @typesafeai thank you! AAC kit is exhausting - forcing a hunt through grids and having no context of what the other person says. I had Jev make a conversationally and context aware decision engine to pick… https://t.co/ESUAJcZ8r9

As promised… I just open-sourced the Jev X (twitter) Market Sentiment Analysis terminal powered by @typesafeai - Ingests 50 - 1,000 live tweets per crypto asset (e.g. BTC, SOL, ETH, etc.) - Scores each tweet for bullishness vs. bearishness using JEV - Combines social sentiment with live funding rates, RSI & volume - Detects short squeeze risks & outputs structured entry/stop/target cards - Built-in SQLite deduplication so you don't burn API credits on repeated searches Powered by Python, CCXT & TypeSafe AI's Jev model. Fork it, tweak it, build on it, or sell it ...but don't forget to enjoy

Built an AI dispatcher for Zammad tickets using Jev (TypeSafe AI). New ticket comes in, Jev decides which team owns it with a confidence score attached. Not a chatbot, a structured decision. Bridge is a tiny Ruby/Sinatra service. Open source, MIT. https://t.co/bXdJpGWHuw #jev

let's start simple. using jev for paragraph-ization: give a block of text and it will add paragraph breaks. I demo it using a text-only youtube client named jevtube. no videos, no distraction, just gets the transcript and adds paragraph breaks. https://t.co/rWuXnxBPUx

started using jev as a judge in @_moodshelf_ and the rankings in semantic search results and items classified by moods, have significantly improved, that too at much lower latency and cost

🔮 Built a real-time click prediction right over the DOM using JEV that feels like a crystal ball. Powered by jev-latest to predict the next clickable element and intent in 345ms per check, highlighting candidate elements on screen. It's extremely cheap to use and helps predict user intention, hitting 83% accuracy in testing. Usages that come to my mind: - Web performance optimization via predictive resource load - UX Research and testing - Progressive UI loading based on intention - Accessibility for reduced motor precision (highlight based on intention allowing to click, for example) - An

Built an SLO-aware inference router using Jev It uses Jev as a typed decision model to select the optimal LLM for each request based on predicted quality, latency, cost, and live backend load Releasing full walkthrough video soon https://t.co/XUjKj9Ryyn

Built jevmetrics — an OTel Collector component using Jev inference to score metric relevance, redundancy, and keep/reduce/drop decisions. Fail-open, with a deterministic policy still in control. https://t.co/1zLhyNjoEp #OpenTelemetry #Observability #Jev #Typesafe

I just had to explore using Jev with my existing Japanese print metadata extraction pipeline (where I use gpt-5.6 luna). Turns out that in some cases I could replace luna completely and save a bunch of money - in others I could augment what I had for higher quality! https://t.co/iSFOGPJjf8

.@jonptaylor recorded a detailed walkthough of Jev vs GPT-5.6 Luna as the "operator" element of a Pipecat speech interface pipeline. GPT-5.6 Luna: - 81.3% command accuracy - 1,008 ms median latency Jev - 92.6% command accuracy - 296 ms median latency A few notes here ... 1) We expected to see a big latency benefit. But the higher accuracy is maybe more interesting. Jev (with a bunch of code wrapped around it) is much better at turning messy transcriptions from a noisy speech environment into correct command structs. The very hand-wavy explanation here is something like: LLM too