What does a jev-ui-test case look like?

A YAML file where the user writes only natural language: an id, a URL, a login, a goal as numbered steps, and expectations with a mode such as and, or or min_pass. A converter skill, /nl-case-author, turns existing functional cases into that shape.

How does Jev fit into the browser loop?

At each step the framework snapshots the visible elements and asks Jev for the operation and target in a single request using speculative fan-out; only the header of the chosen operation's target is executed.

Why can Jev not click an element that does not exist?

Its choice is validated: the selected index must fall inside the candidate ids, the probabilities must normalize, and the pick must be the maximum, or the whole response is rejected and re-sent. Model output can never become a selector, a coordinate or executable JavaScript.

How are pass and fail decided?

Probabilities are evidence, not verdicts. Semantic assertions use Jev's noul probability against thresholds written in pytest, while text_not_contains and url_contains checks are deterministic, and assertions evaluate only on the final page.

What are the measured decision speeds?

The bundled demo report shows a 458 ms median per operation decision and a 919 ms p90, with 30 of 32 decisions under one second; the README labels the older 4-5 s figures as pre-rewrite experience without machine-readable evidence.

What do I need to run the demo tests?

Python 3.12+, uv, Chrome reachable through Browser Harness, and a .env with TYPESAFE_API_KEY and TEXT_MODEL_API_KEY. The default pytest run stays offline at 143 passed and 6 skipped; natural-language cases only run under --nl.

How do I read a failed run afterwards?

The Allure report records candidate counts, the chosen index, operation and target probabilities, decision time and retry counts per step, plus a screen recording and a step timeline; a generated report ships with the repo.

CommunityJev Tools & CLIs

Jev UI Test

A UI testing framework where Jev scores candidate elements instead of generating next-step prose: natural-language YAML cases run through pytest and a browser agent, with Allure reports and a 458 ms median per decision.

See highlights

Open github.com

Stars: 56
Language: Python
License: MIT
Published by: Community
Author: buer2233
Last verified: Oct 2026

buer2233/jev-ui-test

User repository on GitHub

View on GitHub

Jev 决策模型驱动的 UI 自动化测试框架：不生成「下一步做什么」的文字，而是在候选元素里直接打分；一句话写用例，pytest 执行，Allure 出报告，每个操作决策中位 458 ms（UI test automation driven by the Jev decision model — scored decisions, not generated prose; ~458 ms per decision）

Forks: 11
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.Homepage

Tagscommunity python testing browser pytest

Works forJev for Tool calling Jev for Verification Jev for Evaluation

Used inJev for Software & DevTools

Languages: Python
Platforms: Browser
Capabilities: Open sourceRequires API key

Highlights

Cases are natural-language YAML; pytest runs them and Allure reports every decision with probabilities and timing.
Median 458 ms per operation decision, p90 919 ms, with 30 of 32 recorded decisions under one second.
Jev returns operation and target in one request and can only pick from the current element table, so selectors cannot be invented.
Semantic assertions keep the threshold in pytest code: Jev's noul supplies a probability, the verdict stays deterministic.
The repo distributes its own Allure report with screen recording and a step timeline for a real three-case run.

Watch out

MIT, forked from browser-use/jev-ultrafast. Needs Python 3.12+, Chrome through Browser Harness, and both TYPESAFE_API_KEY and TEXT_MODEL_API_KEY; the older 4-5 s figures are the author's prior experience rather than machine-readable evidence, and the offline default run catches natural-language cases only under explicit flags.

Questions about this entry

Short answers drawn from the linked project or page.

Reactions & coverage

Posts, threads, and videos about this entry from around the web.

X: Jev driving real Chrome from Grok Bot

X: jev-browser, a general browser skill

X: A second-hand shopping agent

X: A job crawler that reads careers pages with Jev

More Jev resources like this

surf

635

A CLI browser automation tool with optional Jev semantic mode: Surf observes, Jev selects the next action, Surf validates and acts, Jev checks the goal.

Jev Tools & CLIs#community#browser#cli

Community

nicobailon/surf-cliTypeScript

jev-browse

A Python toolkit that hands browser sub-tasks to Jev instead of the coding agent: one fast_run call lets Jev drive clicks and typing step by step, and the agent gets back a typed outcome or a typed reason to stop.

Jev Integrations & Adapters#community#python#browser

Community

danielnc/jev-browsePython

Translation QA Gate

A localization pre-review gate that pairs source and target strings and judges meaning, terminology, tone, and placeholders as separate questions, so one failing dimension blocks a release without blending scores.

Jev Tools & CLIs#community#python#localization

Community

TradingAgents

111.2k

Multi-agent trading research with optional Jev screening of StockTwits and Reddit posts for company relevance before sentiment analysis runs.

Jev Tools & CLIs#community#python#trading

Community

TauricResearch/TradingAgentsPython

Skill Scanner

2.6k

Cisco's security scanner for agent skills added a System One analyzer: it sends the full skill text to POST /v1/systemone as one advisory Noul question and records the probability, while being deliberately unable to emit a finding or change a severity.

Jev Tools & CLIs#community#security#agent-skills

Community

cisco-ai-defense/skill-scannerPython

pyGPT

A desktop AI assistant with a Jev / System One inline plugin for classification, routing, verification, and scoring without chat-style output.

Jev Tools & CLIs#community#python#desktop

Community

szczyglis-dev/py-gptPython

From the Jev guides

Original write-ups on what this entry is for.

Jev on the Gateways: Vercel, Workers AI, and OpenRouter

Oct 12, 2026

The request shape barely changes across surfaces; the surrounding contract does. When to call the API directly and when a gateway, a Worker binding, or a router entry earns its extra hop.

Jev Tools & CLIs#gateway#vercel#cloudflare

Read article

Exporting Decision Models: GGUF, ONNX, and Device Runtimes

Oct 10, 2026

Taking a decision checkpoint out of Python: what the GGUF and browser WebGPU exports preserve, what they change, and the parity evidence the model cards publish.

Jev Tools & CLIs#open-models#local#onnx

Read article

Records at Rest: Documents, Inboxes, and Risk Queues

Sep 28, 2026

Jev labels a record that already exists. Code moves, flags, or holds it. How the document, mail, and transfer tools keep the action out of the model.

Jev Tools & CLIs#classification#email#security

Read article

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai