agentrun
★ 423▲ 13Parcha-ai/agentrun
A workflow DSL for agents with a Jev package for the steps that are decisions rather than generation.
Recursive Choice over a taxonomy, for picking among more options than a single Choice question allows.
View on GitHub →Hands-on review
A Choice tops out at 255 options; real catalogues are bigger. This walks a taxonomy and asks once per level, so a deep tree becomes a few calls.
Good for
Watch out for
Tested Sep 20, 2026 at 95bff63bd653 · Node 24 in Docker, offline; 14 tests
How we reviewed this: we ran its 14 tests in Docker on Node 24, offline. We made no Jev calls.
jev-tree, by Jalil, solves a real constraint. A Choice question holds at most 255 options; a product catalogue or an industry taxonomy has thousands of leaves. So it walks a JSON tree and asks once per level — partitioning a level that is itself too wide — until it reaches a leaf.
Same stack as the author’s jevlogs: the Vercel AI Gateway, Node 22+, MIT.
The trade in any hierarchical classifier is that mistakes compound: pick the wrong branch at the top and no amount of care below recovers it. Shallow, broad trees fare better than deep narrow ones, and the probabilities at each level are worth keeping if you want to detect a shaky path.
Cost scales with depth, not with catalogue size — which is the whole point.
A small, well-aimed utility. If you’ve hit the option limit and started thinking about two-stage prompts, this is that idea done properly and packaged.
For the same problem inside an application, see how tax-doc-classifier folds five dense form families into a second question.
See how it compares with other tools in Best Jev tools, tested hands-on.
Review updated Sep 20, 2026. Numbers quoted from the project are its author's own; we don't publish our own measurements of Jev.
Parcha-ai/agentrun
A workflow DSL for agents with a Jev package for the steps that are decisions rather than generation.
osolmaz/switch
TypeScript framework for decision-model harnesses: your program offers a short list of next steps and the model picks one at a time until it can answer, with chess, wikirace and rubric-judge examples run on Jev.
disler/ten-levels-of-jev
Thirty worked uses arranged in ten levels, from one if statement to an agent that reaches for Jev itself, with a video walkthrough.
New Jev releases, pricing changes, and the best new projects, once a week. No spam; unsubscribe anytime.
Powered by Buttondown. See our privacy policy.