Case wall
Jev AI examples, page 11
615 real Jev AI builds and public demos, updated daily. Original text, metrics, and a link back to each source.
Last updated
Public Jev AI examples and demos, not official TypeSafe docs. Clients and SDKs are in the Jev GitHub projects. Paths for jev-latest, Vercel, and OpenRouter are in the September 2026 channel map.
Agent-skill repositories are listed on Jev skills on GitHub.
OpenRouter model ids recorded for Jev AI are on Jev AI on OpenRouter. Rubric and eval posts are on Jev as a judge. How the company and the model fit together is on TypeSafe AI. The pricing calculator uses the published input rate.
No posts match these filters.
@b1gdan uses Jev on an inbox: 263 emails sorted into piles in 2.0 seconds, labeled client, lead, needs a reply today, or needs acting on.
I'm using Jev with my inbox, and it sorted 263 emails into piles in 2.0 seconds, then tells me if each one is a client, a lead, needs a reply today or needs act…

@felipe_rohde built MenuBlip: a digital menu that understands orders in any language and recommends dishes in real time on Jev.
I built MenuBlip, a digital menu that understands orders in any language and recommends dishes in real time. It runs on Jev, the first System One model by @type…

@araskavakli built a contact-form screener with Jev and Claude Haiku: legitimate enquiries pass, the rest wait, about $1 a month, no captcha.
A client's contact form was being used to phish and harass their team. We built a screener using Jev and Claude Haiku. They read every message first. Legitimate…

@SaadMRP: Jev diagnoses SREGym-Lite faults; LLM review lifts pass rate 78.1%→89.5% across 21 faults, ~38s avg.
1/8 In our latest experiment, we let Jev, a decision model, investigate system failures and asked an LLM to check its diagnosis. Across 21 SREGym-Lite faults, r…

@PeteCapeCod ships npm dev-orchestrator: Jev/OAI Decisions/CF Clef drive Codex CLI or Claude Code instead of expensive reasoning tokens.
I've been working on a new thing! It's a cli that uses other clis, basically you can use Jev (or OAI decisions, or CF's Clef/flash) to make calls either codex c…
@parkerbrowndev hooks Jev before tool calls in a Pi/Harbor setup with Haiku; reports better safety, half the tokens, higher API $ from broken caching.
Jev pre-tool call router experiment pass 1 is in! Design - Haiku (fast) model - Pi harness - Harbor eval - Jev gets hooked in BEFORE tool calls Result: - MUCH b…

@jurlycat cites Vercel’s October AI Gateway report: Jev reached 35.3% of paid teams in 14 days with 0.27% of spend and strong week-two return.
Jev in Vercel’s October AI Gateway report: • 35.3% of paid teams used it in its first 14 days, the fastest adoption in Gateway history. Sonnet 5.5 reached 18.1%…
@techemmyleke built an open-source scene finder: describe a movie moment and Jev locates it; Windows and Apple silicon.
👀 One use case for jev is being able to find scenes in movies that you just cannot remember the timestamps. Built scene finder to demo that concept. It lets you…

@sinkush: a local hybrid pass takes the top 30, then Jev judges each hit in parallel in about 0.3 s for a tenth of a cent.
Semantic search isn't relevance search. I first built agentic, file-system-style retrieval: accurate, but ~30 s per question. Now a local hybrid pass finds the…
@dave8172 uses Jev in an affiliate-terms refresh pipeline to validate terms and extract only when a page changed, with per-page time and cost versus Haiku and Opus.
Added @typesafeai Jev to speed up data check on webpages. I run an affiliate program terms directory and for its refresh pipeline, which regularly has to update…
@shannholmberg's six-step guide: use LLMs for planning/coding and JEV for yes/no decisions (~$0.04 per 1,000 vs dollars on GPT/Claude).
how to add a decision model to your AI workflow 1,000 yes/no decisions can cost around: ~$10 on GPT-6 Astra ~$4 on Claude Opus 5.5 ~$0.04 on JEV use language mo…
@trevin demo: name an object in a picture; Jev 1.14 draws a bounding box and shows latency and cost (demos.trev.in/hotdog).
Built a small demo of @typesafeai Jev 1.14 which adds image support: Where is Hotdog? 🌭 Name something in the picture and Jev tries to find it, drawing a boundi…

@danielhayesmith wires TypeSafe Jev to analyze after each sentence on a custom phone keyboard he built.
OK, I added @typesafeai Jev to analyze after each sentence in this new keyboard I've built. Pretty wild!
@kj4dh4v built the Seefood app with TypeSafe Jev image and @googlegemma. It goes beyond hotdog or not hotdog. Try seefood.verygoodproduct.company.
Built the Seefood app using @typesafeai Jev image and @googlegemma. It goes beyond hotdog and not hotdog now. Try it: https://seefood.verygoodproduct.company
@typakon4 shipped jev-layer 0.3.0 with Jev 1.14 screenshot input, visual grounding, and opt-in desktop action advice for agent harnesses.
Shipped jev-layer 0.3.0 with Jev 1.14: screenshot input, visual grounding and opt-in desktop action advice for agent harnesses. Jev recommends. Your host keeps…
@anarchyco uses Jev-style decisions (on-device Tev1) to detect basketball regions of interest, compared against YOLO + OC-SORT heuristics in a video.
Trying Jev but in a different way.... Using it now to detect Region of Interest in a basketball game and also running it on-device with Tev1 implementation inst…
@realnribal's pre-execution risk gate: Jev scores rm -rf ./node_modules at ~50% and blocks rm -rf / at 99%.
I Built a pre-execution risk gate for AI agents using Jev (TypeSafe AI). rm -rf ./node_modules → ~50% risk, asks first rm -rf / → 99% risk, blocked instantly Sa…
@akshay_pachaar put Jev in TrueForge for client follow-up matching: as accurate as Opus on 40 requests at 150x lower decision cost.
i just put Jev inside an open agent harness. (150x cheaper decision-making than LLM) agents burn a lot of tokens on decisions that are really just picking one o…

@mentiss_ai ships TypeSafe Jev in Mentiss for votes/targets; 10-player test cut model costs over 60% vs older games.
I just shipped Jev from @typesafeai in Mentiss. It handles votes, targets and other non-speaking decisions; your chosen model still does the talking. In my 10-p…
@dfinke: screenshot→spreadsheet in 11s; Jev finds the table (skips chart), then LLM+ImportExcel via one PowerShell line.
Screenshot → spreadsheet in 11 seconds. One line of #PowerShell. @typesafeai Jev finds the table (skips the chart), crop → LLM extracts the rows, ImportExcel sh…
@CaineArdayfio demos Thud: Mobile Use plus Jev audio that maps spoken commands to page actions without transcription.
Mobile Use + Jev for audio = ULTRAFAST⚡️ 10x faster and 20x cheaper than Gemini Live. It's our Audio Action model, Thud. Thud listens and then does. It has a pr…

@gbolagadehq compares Jev 1.14, OpenAI Decisions API, and Opus 5.5 on five public tests including 1,999 product photos.
https://x.com/i/article/2112290824528576512

@twid put Jev 1.14 on a doorbell→Discord pipeline: person-in-frame checks hit 23/23 in ~200ms, then Box crops for captions.
Tried @typesafeai’s jev 1.14 preview on my doorbell cam’s alert-to-Discord pipeline. My old setup: Gemma 4 26B describes the frame, then DeepSeek V4.1 Flash wri…
@kmelve wired multimodal Jev box answers into Sanity so drafts auto-set image hotspot and crop from a text description.
Jev got 👀 📐🖼️ now! Multimodal Jev and a new bounding box type from @typesafeai launched today as early access aka "dreamer mode" aka `typesafe_danger_zone`. Too…