Each entry here comes with a moving picture of the thing working: a screen recording attached to an X post, a YouTube video, or an animated GIF from a project's README. Most are X posts from the days after launch, so the view counts reward what spread early more than what works best. The cards show a still. Open an entry to watch the original, and check its date before you trust a speed claim.
201 demos, most viewed first
Showing 1 to 48 of 201
The founder's launch post, around 76k likes, arguing that RLCD-trained decision models are a shorter path to economic value than chat models. Submitted separately to HN as item 49716682.
Diogo Almeida (@CompleteSkeptic)OpenDiogo Almeida's Jev launch post (X) on x.com39M viewsSpeed 20-200x fasterCost 40-400x cheaper
An independent MLX port of Laya for Apple Silicon, 13.4ms median for a short English decision and 7.4ms on the multilingual checkpoint, with a Snake demo that plays at up to 75 moves per second under 1GB of memory.
mizorewwwOpenLaya-MLX on github.com4.2M viewsShort decision 13.4 ms P50Snake demo 75.4 moves/s
A video demo of copy and paste made 'smart' using Jev from TypeSafe, with no further detail on the mechanism given in the post text.
A demo breaking down 724 live ads from 37 brands in 40 seconds for 9 cents, covering hook, format, offer, CTA, funnel stage and landing-page mismatch for each ad.
Matthew Berman (@TheMattBerman)OpenBreaking down 724 ads from 37 brands with Jev (X) on x.com1.1M viewsSpeed 724 ads/40sCost 9 cents
A video demo has Jev read 586 pages of a site in 45.1 seconds and rebuild the internal link map, placing 584 links and refusing 139 pages where nothing fit, for $0.21 total. In the same time, Claude Opus 5 processed only 21 of the same pages for $1.43.
borja (@borjafat)OpenAn internal-link SEO audit (X) on x.com827k viewsPages read 586 in 45.1sCost $0.21
A video demo rebuilds a Tesla Full Self-Driving-style simulator controlled by Jev, built in under an hour.
Company launch post announcing early access to Jev.
Open source skills that read a morning's news feed with Jev and tell a list of brands which stories to act on. In a demo run it read 384 headlines and matched them to 15 brands in 24.9 seconds for $0.19, against 4 of 384 for $0.77 on Claude Opus 5 over the same feed.
Elvis (@elvissun)Opennewsjack on newsjack.sh661k viewsCost $0.19vs Opus/headline ~390x cheaper
A video demo of voice-controlled computer use on a Mac built with Jev: dictating a command such as opening the Notes app triggers the action before the sentence finishes.
A video demo of Jev playing Super Mario Bros. in real time, using its fast inference and structured output for moment-to-moment decisions.
Deploys DiffusionGemma-Jev (djev) as a Jev API-compatible endpoint on Cloud Run with one RTX PRO 6000 GPU, scaling to zero when idle. The Gemma team's video demo reports 35 to 60 ms single-step latency at roughly $3 an hour.
Google Gemma (@googlegemma)594k viewsOpendjev-run: DiffusionGemma-Jev on Google Cloud Run on github.com
A video demo of Jev V13 playing 5+0 blitz chess against Fable 5.1 and GPT-6 Astra, one move per API call, answering in about 2.6 seconds per move. Fable was ahead on material but lost on time; Astra mated Jev in 18 moves.
AI/ML API (@aimlapi)OpenBlitz chess against frontier models (X) on x.com559k viewsMove latency ~2.6 svs Astra mated in 18 moves
A video demo of Jev generating platformer game levels in real time, using fast structured output aimed at game development.
A short video demo asking whether Jev understands colour, posted by generative artist Matt DesLauriers, who calls the model a leap forward for building new UX and UI paradigms on top of language models.
A video demo of jev(), a PostgreSQL extension that filters rows with a natural-language condition inside a WHERE clause with no index or embeddings. It judged 129 rows in about 1 second for $0.0009, and 6 milliseconds on a cached rerun.
Zachi (@iam_zachi)OpenNatural-language queries inside PostgreSQL (X) on x.com478k viewsFirst run 129 rows/~1sCost $0.0009
Given 700 high-intent leads and personalized outreach messages, Jev predicted how each message would perform, assigned a confidence score and flagged lead-message mismatches in 40 seconds for $0.09.
Romà n (@romanbuildsaas)OpenScoring 700 leads and outreach messages with Jev (X) on x.com476k viewsSpeed 700 in 40sCost $0.09
A video demo of an email classifier built with Jev, tested on 1,500 of the author's own emails.
Demos a launcher that ranks results by alias, fuzzy match and habit; Jev reads typed intent so a query like 'the pdf I just downloaded' puts the newest matching file on top within about 100 ms.
nader dabit (@dabit3)OpenA predictive launcher powered by Jev (X) on x.com429k viewsResponse time ~100 ms
A video demo of Jev drawing pictures by predicting every pixel in parallel.
A video demo of a massively parallel, browser-based adversarial testing suite built with Jev that tries to break each software release, at a cost described as pennies per run.
Walks through the Choice, Score and Noul primitives live in the TypeSafe playground, with speed and cost numbers and their caveats.
A video demo controls a browser in real time by voice: the spoken transcript goes to Jev, which returns probabilities in about 300 milliseconds, and the browser acts on the result for roughly $0.0002 per decision.
Moritz Kremb (@moritzkremb)OpenReal-time voice control for a browser (X) on x.com368k viewsResponse time ~300 msCost per decision $0.0002
Describes an SEO/GEO agency workflow, auditing Search Console, PostHog/Mixpanel and AI citation data, that used to cost about $250 per client; each step now runs roughly 30x faster with Jev, cutting the total cost by 90%.
Ira Bodnar (@irabukht)OpenAn SEO/GEO agency workflow that cut costs 90% with Jev (X) on x.com366k viewsCost cut 90%Prior cost/client ~$250
A demo of Jev playing Subway Surfers at superhuman speed, including 50 games running at once, for under a cent total.
A video demo shows Jev returning 14 typed probability checks per pull request diff, such as hardcoded secrets and SQL injection, which code turns into a block, review, or merge verdict. Checks scoring between 0.35 and 0.65 confidence are escalated to a human or a larger model.
Paolo Rosson (@redp314)OpenFourteen checks for every pull request (X) on x.com357k viewsCost per PR $0.00007vs Opus, 1k PRs 7c vs $14.50
A video demo of Jev controlling all 4 characters in a Smash Bros match against itself, picking each next move within a fraction of a second; the match used over 22 million tokens for a few cents.
Mau Baron (@maubaron)OpenJev playing all 4 characters in a Smash Bros match (X) on x.com311k viewsTokens used 22M+
A video demo shows Jev classifying 500 emails in a few seconds for a total cost of 3.5 cents.
Riley Brown (@rileybrown)OpenClassifying 500 emails in seconds (X) on x.com296k viewsCost 3.5 cents
A video demo of computer-use automation built with Jev, said to run 155x cheaper and about 20x faster than Opus 5 while generalizing across operating systems.
aaron (@awlevin)OpenComputer use across operating systems (X) on x.com292k viewsvs Opus 5 155x cheaperSpeed vs Opus 5 ~20x faster
Hassan used Jev to sort 1,018 AI research papers into one of 24 topics for $0.08 in total, with a 256 ms median latency per paper, after DeepSeek V4 Flash wrote the summaries. He was still running evals before swapping them in for the topics on 1kpapers.com.
Hassan (@nutlope)Open1kpapers, AI research papers classified by Jev on 1kpapers.com286k viewsCost $0.08Latency 256 ms median
A video demo of predictive spreadsheets where typing a column header such as Urgency prompts Jev to rate each row on a scale from no follow-up needed to urgent in about 100 milliseconds.
nader dabit (@dabit3)OpenSpreadsheets that read intent (X) on x.com276k viewsResponse time ~100 ms
CJ from Syntax explains what Jev is, how it works, and walks through several use cases and demos. 17 minutes.
A TypeScript classifier that sends each PDF page's text to Jev as a choice over 261 IRS forms and 7 page kinds, with a second call only for five corporate forms and their schedules. Its eval reports 0 wrong pages on 314 filled TaxCalcBench forms at about $0.001 per page, 34x cheaper and 6x faster than the Sonnet pipeline it replaced.
Nakshatra Saxena (@nedwize)Openkyotofin/tax-doc-classifier on github.com241k viewsCost per page $0.001vs prior setup 34x cheaper
Laya, running locally on Apple Silicon via MLX, and Jev, over TypeSafe's hosted API, play Chrome's T-Rex game side by side on a shared obstacle course, with response times, survival streaks and crash replays recorded.
A video demo of Mac computer use driven by Jev with no screenshots sent to an LLM: a local CoreML model segments on-screen elements, on-device OCR reads their labels, and Jev picks which one to click in a loop, at about 90ms per decision.
Milind S (@milindlabs)OpenMac computer use with local perception (X) on x.com239k viewsPer decision ~90ms
Free SuperX tool that scores a draft X post from 0 to 100 against a normal post from an account your size, with lists of what helped and what hurt and a simulated feed preview. Rob Hallam's launch post says each draft gets 61 Jev questions in about a second, fitted on 9,481 posts from 207 creators.
Rob Hallam (@robj3d3)OpenTweet Tester: Preview & Score a Post Before You Send It on superx.so238k viewsPer post ~1s, $0.0004Accuracy 2 in 3
A video demo of a Rubik's Cube solving itself with Jev: a standard beginner method is coded in, and at each step Jev picks which case the cube is in while code checks the pick. It took 94 moves rather than the 22-move optimal solution, using about 4 seconds of model time.
Paolo Rosson (@redp314)OpenA Rubik's Cube that solves itself (X) on x.com207k viewsMoves used 94 vs 22Model time ~4s total
A free tool where Jev reads 3 days of posts from a chosen niche, asks 8 questions per post to filter out bait and hidden ads, and weighs like/reply/repost ratios, running in about 2 seconds for $0.007.
Rob Hallam (@robj3d3)OpenA feed filter that scores posts by niche with Jev (X) on x.com176k viewsPer run ~2s, $0.007Questions/post 8
A simulation of the X ranking algorithm built with Jev, using real ranking weights to simulate a post's virality against a global feed of other users.
A video demo classifies 1,891 ads from a competitor's ad library in 19 seconds for $0.12, tagging each by customer journey step and ad style.
Stav Zilbershtein (@mightyking)OpenMaxfusion competitor-ad research (X) on x.com171k views1,891 ads 19s, $0.12
A video demo building a small language model out of a classifier: Jev answers 29 yes/no questions per character, the highest-probability answer is appended, and the updated text is fed back in for the next character.
vogel (@ryanvogel)OpenWriting one character at a time (X) on x.com167k viewsQuestions/char 29
A video demo evaluates a candidate profile against 400 companies in 12 seconds, ranking job matches with a confidence score and flagging candidate-job mismatches, for a total cost of $0.0005.
Sarvagya Kulshreshtha (@sarvagya_kul)OpenJob and candidate matching (X) on x.com150k views400 companies 12 sTotal cost $0.0005
A video demo uses Jev for incident triage in Box: it answers whether an incident is customer-facing and how severe it is, then routes the file into Escalate, Monitor, or Review folders and sets a metadata template with the result. Incidents below a confidence threshold go to Review for a human instead of routing automatically.
A video demo of a macOS app that monitors the Downloads folder with Jev: it checks whether a downloaded file is an invoice, then moves matching files into a folder with a corrected filename, using no other LLM calls.
Marcel Pociot 🧪 (@marcelpociot)139k viewsOpenAutomatically organize downloaded files (X) on x.com
A video demo of intent-based email search in Gmail built with Jev, with the author suggesting a large inbox would first narrow results with semantic search or embeddings before Jev ranks them.
A follow-up demo where Jev scrolls 723 ads as 30 simulated buyer personalities, producing 21,690 stop-or-scroll decisions for 22 cents.
Matthew Berman (@TheMattBerman)OpenScoring 723 ads against 30 simulated buyers with Jev (X) on x.com131k viewsDecisions 21,690Total cost 22 cents
A video demo backtests Jev trading the Danish stock market across all 239 trading days of 2025, using sentiment derived from market data, news articles, Wikipedia, and Google Trends, at a cost of 8.1 million tokens for $0.32.
Tommy Jepsen (@tommy_jepsen)OpenDanish stock-market backtest (X) on x.com125k viewsToken cost 8.1M tok, $0.32
Jev filters and selects creative assets, which DeepSeek and Higgsfield then turn into ad creatives, shown as a video demo from Higgsfield AI.
Higgsfield AI 🧩 (@higgsfield_ai)123k viewsOpenFiltering creative assets for ad generation with Jev (X) on x.com
Local-first macOS coding workspace built in Rust with a local Laya selector and optional hosted Jev, routing fresh coding tasks and validating each chosen route on the host.
Avid (@Av1dlive)Opencodejunkie99/keel on github.com117k viewsCost/time cut ~80%
Nothing matches those filters
Clear a filter to widen the search, or send us the thing you were looking for.
