561 catalog entries · 11 categories
Jev AI tools categories
Every project in this directory comes from the community-maintained awesome-jev list and is sorted into the categories below. Pick a category to see the projects, what each one does, and which Jev composition patterns pair with it.
This site is an independent third-party directory of Jev AI tools and is not affiliated with, endorsed by, or operated by TypeSafe AI.
Official
First-party Jev resources published by TypeSafe AI: the company site, documentation, quick start, HTTP API reference, and community Discord. Everything listed here is maintained upstream, not by this directory.
SDKs & clients
Client libraries that wrap the Jev / System One API so you can ask typed questions from your own stack instead of hand-rolling HTTP calls. The official TypeScript and Python SDKs are listed first, followed by community-maintained clients in other languages.
Integrations
Glue code that puts Jev decisions inside tools you already run — coding agents, agent frameworks, MCP servers, validation libraries, and databases. These projects are community-maintained and typically require your own Jev API key.
Agent tooling
Tools that make agents decide faster. The common shape is a reflex router: keep the agent state, ask a few typed questions — next action, needs human, complexity — and branch in code, rather than paying multi-second generative latency per step.
Browser & computer use
Web and desktop agents click many times per task, so per-step cost and latency dominate. Projects here use Jev to pick a target from a candidate list, flag blocking overlays, or choose the next screen action — keeping each step typed and cheap.
Applications
Complete applications rather than libraries. These projects show what a Jev-backed product looks like end to end: where the typed decision sits in the request path, what gets cached, and how the answer surfaces in the interface.
Games & simulations
Games are a natural fit for typed decisions — the answer space is usually a small enum or a score, and latency is visible to the player. Projects here use Jev to drive NPC behavior, branch simulations, or generate study material on the fly.
Demos & playgrounds
Hosted experiments you can open and use without installing anything. Good for building intuition about how a constrained answer looks and how quickly it comes back.
Benchmarks & research
Evaluation material — measured latency, task accuracy, and reproductions of published claims. Treat these as third-party work with their own methodology, and read the setup notes before quoting any number.
Other lists
Curated lists maintained by other people. Overlap with this catalog varies, and some cover broader AI-agent territory rather than Jev specifically.
Articles & threads
Longer-form and conversational material: blog posts, engineering notes, and social threads where builders describe what they tried. Useful for judging where the approach fits and where it does not.