Projects
Things people built, from weekend demos to products already running.
323 public things people have built, written or argued about since Jev shipped: code, tools, official cookbooks, videos, articles, threads, lists, and the open-source alternatives that cover the same ground. If something is missing, use the submission form.
Things people built, from weekend demos to products already running.
SDKs, wrappers, MCP servers and integrations that call the API for you.
Official worked examples and patterns, straight from the vendor docs.
Walkthroughs, launch coverage and explainers, in several languages.
Write-ups, reviews and analysis, mostly from outside the vendor.
Threads and posts where people argue about what this actually is.
Awesome lists and roundups that collect everything else.
Open-source models and libraries that cover similar ground.
The entries worth opening first: official material, the tools people actually install, and the demos that show the model doing something.
Browser Use's fast browser agent where Jev picks the operation and the DOM element in a single request; a small LLM is only invoked to write text when typing is required. Demoed on a Google Flights search.
Attempt to run a Jev-style System One decision model locally on a single RTX 3090. Companion site at openjev.com.
Open reimplementation that trains a small model to choose among a changing list of text options, emitting one probability per option in a single forward pass. Ships Doom, chess and Wikispeedia demos.
Makes one AI trade decision every Monad block on the Kuru MON-USDC market using Jev. Live at jev-trader.vercel.app.
Independent community collection of use cases, patterns, prompts and starter code for Jev, explicitly not official. The 2023 creation date means the repo was renamed from an older project, so its 412 stars are not Jev-era traction.
Official agent skill for Claude Code, Codex and other agent environments; installs via claude plugin marketplace add typesafe-ai/skills or npx skills add typesafe-ai/skills. MIT licensed.
Proof-of-concept MCP server for Jev that lets Claude Code, Claude Desktop and Codex call the model and receive probabilities they can branch on. The most-starred Jev MCP server found.
Walks through the Choice, Score and Noul primitives live in the TypeSafe playground, with speed and cost numbers and their caveats. The most-watched Jev video found.
CJ from Syntax explains what Jev is, how it works, and walks through several use cases and demos. 17 minutes.
AI News daily roundup leading with the Jev launch, summarising RLCD and TypeSafe's 20-200x speed and 40-400x cost claims. Frames Jev as part of a broader move toward task-specialised models.
Argues Jev's speed comes mainly from constrained single-token generation and aggressive prefill, techniques reproducible with existing LLMs, rather than a new architecture. Still treats very fast structured output as a new primitive, while doubting Jev matches frontier models on raw intelligence.
Vercel CTO reports Jev saturated an existing classifier eval that had used Gemini 2.5 Flash Lite and ran about 6x faster. Original post; TypeSafe's quote-tweet is a separate item.
The same entries, sorted by the job they do.
The twelve most recent additions.
Send the link. Entries are checked by hand and added most days.