Codex and Jev: How to Speed Up Browser Automation by 5-10x
Codex and Jev: How to Speed Up Browser Automation by 5-10x
Integrating Codex with Jev allows delegating routine browser actions, reducing LLM calls and speeding up workflows by 5-10 times.
We recommend trying the combination of Codex and Jev. Together they speed up browser interaction by 5–10 times.
Usually, when Codex works with a page on its own, it has to read the DOM, reason, and call a tool at each step. Each such cycle adds latency.
Now, actions like clicks, navigations, switches, and scrolling can be delegated to Jev, leaving Codex to handle reasoning, text input, and result verification via image.
Key advantages:
Jev executes actions sequentially, while Codex connects only when something needs to be entered or an image needs to be checked
The fewer requests to the model, the lower the latency and cost
How to set up:
- Prepare Codex with a browser/CUA, Node.js 22+, and a TypeSafe API key
- Install the Skill: pass the Codex repository or run
npm run install:codex -- --env-file .env - Load
jev-useso that Jev takes over clicks, navigations, and scrolling - When a decision needs to be made or text needs to be entered, control returns to Codex. After that, Jev continues execution
- Test everything on a real product page or in an automated test
For those who independently develop products and run tests through the browser every day, it is worth trying.
The Skill is open .com/vlad-terin/jev-use](http://github.com/vlad-terin/jev-use)
Why it matters
AnalysisThis approach optimizes AI resource usage by reducing the number of expensive and slow LLM calls for routine browser actions.
Discuss in community
Share your questions and insights with developers
Related in Ecosystem
Practical guides, core terminology, and agent skills
Jev and System One AI: A Complete Practical Guide to the Decision Model from TypeSafe
A deep engineering breakdown of the flagship Jev model: System One architecture versus LLMs, Choice/Score/Noul primitives, request parallelism, confidence mathematics, video use cases, code, and an overview of local open-weight alternatives (Laya, GLiNER2.5, CLM-8B).
ChatGPT Updates: Plugin Development and Codex Workflows
Comprehensive engineering guide to creating, packaging, testing, and submitting plugins for ChatGPT and Codex: manifest specification, MCP servers, lifecycle hooks, custom marketplaces, and OpenAI review.
OpenAI Codex AI agent damaged developer's server
A developer asked OpenAI Codex to reboot a server, but the AI agent disrupted disks, systemd, and containers, ran unrequested scripts, and caused more than $10,000 in damages.
AI Agents (Autonomous Agents)
Software systems based on LLMs that can autonomously perceive the state of the environment, decompose complex goals, invoke external tools, and iteratively correct their own mistakes.