Does the runner launch or clean up my browser for me?

No. You launch Chromium, open the page, navigate, and close it. The runner only borrows the Page you pass in, so auth, interception, and test isolation stay in your setup code.

Do I really need verifyCompletion for side-effecting flows?

Yes when clicks or submissions matter. Without it DONE means the model judged the task done, not that the visible outcome exists, so check for expected text or state explicitly.

Where do API keys go, and what endpoint does it call?

Keys live in the environment, never in YAML. It defaults to OpenRouter Decisions with a Jev model, keeps a compatibility key fallback, and allows base URL and model overrides for compatible gateways.

What does it send or log that I should worry about?

It strips query strings and fragments from URLs in observations, excludes password and file inputs, and never logs keys or auth headers. You still own request interception and domain-specific write guards.

How do I install the runner, and which Node versions work?

Run pnpm add @chlrc/midscene-jev-runner @midscene/test playwright. Node.js ^20.19.0, ^22.12.0 or >=24.0.0 is required.

Which environment variables configure Jev?

Set OPENROUTER_API_KEY; the runner calls OpenRouter Decisions with ~typesafe/jev-latest by default. MIDSCENE_JEV_BASE_URL and MIDSCENE_JEV_MODEL_NAME select compatible gateways, and a TYPE_TEXT action needs the MIDSCENE_JEV_TEXT_* or MIDSCENE_MODEL_* variables.

What does the runner strip or exclude before observing?

It strips query strings and fragments from URLs sent in observations and excludes password and file inputs from the action space, and it never logs API keys or authorization headers.

CommunityJev Demos & Experiments

Midscene Jev Runner

Community runner that adds a strict jevAct step to Midscene Test plus a direct runJev helper for Playwright scripts, with caller-owned browser control and optional completion checks.

See highlights

Open github.com

Stars: 11
Language: TypeScript
License: MIT
Published by: Community
Author: KiritoKing
Last verified: Oct 2026

KiritoKing/midscene-jev-runner

User repository on GitHub

View on GitHub

Community-maintained JEV runner integration for Midscene Test

Forks: 0
Last pushed: Updated Sep 2026

Repo stats from the GitHub API, cached Oct 2026.

About this project

Community-maintained JEV runner integration for Midscene Test. This project is independent from the Midscene repository and is not an official Midscene package. runJev(page, options) for direct Playwright use. createJevNodes(options) for registering a strict jevAct node in a Midscene Test project. A caller-owned browser lifecycle: the runner does not create, navigate, route, or close the supplied Playwright Page. Optional completion verification and structured observer events.

Quoted from the project's README. Read the full README.

Tagscommunity awesomejev catalog-import

Works forJev for Classification

Used inJev for Software & DevTools

Languages: TypeScript

Highlights

Registers a strict jevAct node via createJevNodes and supports direct calls with runJev(page, options).
Leaves browser ownership with the caller: it never creates, navigates, routes, or closes the Page.
Recommends verifyCompletion for side effects, since DONE alone means the model stopped, not success.
Strips URL query strings and fragments from observations and excludes password and file inputs.

Watch out

Needs OpenRouter or compatible gateway credentials in the environment, and the caller owns auth, navigation, and cleanup. Copy reviewed against the repository README on 2026-10-04, not from running it.

Questions about this entry

Short answers drawn from the linked project or page.

More Jev resources like this

laya-server

Self-hosted Docker service and web UI that packages Laya structured-decision inference behind a compatible System One endpoint.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

1Panel-dev/laya-serverPython

jot

General-purpose System One agent where Jev chooses tools and arguments while Jot executes them and loops until it can reply.

Jev Demos & Experiments#community#awesomejev#catalog-import

Community

runta-dev/jotTypeScript

jev-reflex-autonomy-lab

Browser drone-fleet lab where TypeSafe Jev steers each drone as fast reflexes and an optional planner offers one-use strategy advice when confidence drops, without pausing control.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

khordoo/jev-reflex-autonomy-labTypeScript

x-scanner

Chrome extension that sends each X post to Jev with six typed questions in one call near the viewport, then renders a behavior chip under the post plus a session cost and latency panel.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

oso95/x-scannerTypeScript

lkclean

Manifest V3 Chrome extension in TypeScript that sends each LinkedIn post to Jev and collapses bait, self-promo, off-topic, and sponsored posts into a one-line bar with reasons and scores.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

stefw/lkcleanTypeScript

flue-jev-demo

Standalone Flue example that routes support chat through Cloudflare Workers AI, using gpt-4o-mini for chat and Jev for typed intent and urgency routing.

Jev Demos & Experiments#community#awesomejev#catalog-import

Community

matthewp/flue-jev-demoTypeScript

From the Jev guides

Original write-ups on what this entry is for.

Per-Tick Decisions in Games

Sep 28, 2026

A game tick is a state plus a legal action list. How Clash Royale, Minecraft, and the earlier Mario, Pokémon, StarCraft, and drone loops ask Jev one step at a time.

Jev Demos & Experiments#gaming#agents#control

Read article

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai