Skip to content
Metatext Decide — typed decisions for your agents, with a probability on every answerGet early access
Metatext
Sign Up

CodexQA Jev Browser

CodexQA browser skill that indexes in-page controls and has Jev pick the operation and target each step, with no screenshots needed.

Agent skillbeta
View on GitHub ↗

Overview

CodexQA Jev Browser is a skill in CodexQA, a set of local agent skills that check whether code is actually good after it is written. Rather than sending a screenshot to a vision model each step and clicking coordinates, it builds an index of controls inside the page and treats visible page evidence as the result, with Playwright Chromium executing. Replay, goal runs, case generation and site exploration share the index, and actions target stable data attributes so layout changes do not break clicks.

With a TypeSafe key, each step is a /systemone questionnaire: which operation and which observed target, and the same channel judges whether the action's effect appeared on the next page. Typing uses phrases already present in the goal. A knowledge base supplies site-specific facts Jev does not know. The README quotes TypeSafe's published speed and cost comparisons but states the repository has no benchmark against vision GUI models. Docs are in English and Chinese.

What the decision model decides

TypeQuestionAnswers
ChoiceWhich operation and which observed target should be used next?indexed controls
Yes / noDid that action's effect show up on the next page?yes · no

How it works

  1. Index controls in the live page
  2. Jev picks operation and target
  3. Playwright acts on the indexed node
  4. Jev checks the effect on the next observation
  5. Use site knowledge notes for domain facts

Strengths

  • No vision tokens per step
  • Stable targets resist layout changes
  • Shared index across replay and exploration

Caveats

  • No own benchmark against vision models
  • Typing limited to phrases in the goal

Getting started

Install from the CodexQA catalog or this repository and set TYPESAFE_API_KEY.

At a glance

Type
Agent skill
Status
beta
Runtime
agent skill + Playwright
Has code
yes

Use cases and integrations

How curated lists describe it

  • Jev Browser — indexed browser automation. Jev chooses the control, Playwright acts. A CodexQA skill.

    hellogumbo/awesome-jev

  • Browser automation that indexes the controls inside the page and has Jev choose one, instead of sending a screenshot to a vision model on every step.

    MrJev/awesome-jev

Frequently asked questions

Does it use screenshots?

No, it uses an in-page control index.

What does Jev decide?

The operation and target per step, and whether the action took effect.

Is there a benchmark against vision agents?

No, the README says none exists in this repository.

Related projects

  • jev-browser-qa

    Browser QA where Playwright drives and records while Jev judges assertions, matches controls to intent and picks goal-directed clicks.

    Agent skillbetaTypeScript
  • public-browser

    MCP server and libraries that drive Chrome over CDP and hand a decision model like Jev a bounded menu of page actions.

    ProjectproductionTypeScript
  • Jev Browser★ 335

    Headless browser automation via MCP server, CLI or library: Jev picks one action per step and scores goal-met and stuck probabilities.

    MCP serverbetaTypeScript
  • AskJev

    Claude Desktop MCP extension that drives Brave or Chrome with Jev choosing on-page actions and a guard freezing irreversible clicks.

    MCP serverbetaTypeScript
  • BrowserPaw

    Chrome MCP server with 50 tools that drives your everyday browser without losing logins, where Jev or a local decider picks each step.

    MCP serverbeta
  • CloakBrowser-Agent

    Stealth browser agent where Jev decides each step in about 0.3 s and CloakBrowser carries it out; MCP server, CLI and Python API.

    MCP serverbetaPython