Codi Agent Opens Chrome with Its Own Browser Profile

Codi Agent can open an installed copy of Google Chrome and automatically create its own browser profile named Codi Agent. It uses that profile for browsing and web research, separate from your everyday Chrome profiles. This is part of Codi Agent, Codiosity’s desktop work assistant with optional AI and skills.

You can ask Codi to read pages, collect information from several sources, or work through a multi-step web task. The Chrome window remains visible while in use, so you can see the page being worked on and step in when a site needs your help.

What does a separate Chrome profile mean?

A browser profile separates Chrome sessions and settings. Codi prepares its own profile when the browser feature is used, then reuses it for later work. The Codi Agent name helps you recognize the window being used by the app.

Creating this profile does not require a new Google account. To access a site that needs an account, log in yourself in Codi’s Chrome window. Sessions can remain in that profile according to the site’s rules; the site may still ask you to log in again.

Browsing and research serve different tasks

• Chrome research focuses on opening and reading sources, summarizing information, and including references for the material collected.
• The browser agent can follow a multi-step flow: open pages, follow links, scroll, read, and fill parts of forms according to the task and permissions.
• The same profile can be reused, allowing later work to use browser sessions you have already prepared.

Make the request specific: what information you need, which sources matter, and how you want the result presented. For research, ask for references so you can check important details on the original pages.

An example request to start with

Include a site address when you know the starting page. Specify whether you want information read or a change made. Describing the goal and action boundaries makes the result you need to review clearer.

Prepare Chrome, AI, and an internet connection

The browser feature needs Chrome already installed, an internet connection, and suitable AI or provider configuration in Codi Agent. Preparing a profile does not install Chrome automatically. Review feature choices and permissions in the app version you use before starting.

Users handle logins, verification codes, and CAPTCHAs directly. After completing them in the Chrome window, you can ask Codi to continue. Page changes or site restrictions may require adjusted instructions.

Actions follow your permission settings

In version 1.1.45, browser requests from the app can start without an initial confirmation. Your saved permission preferences still apply; requests from Telegram still require approval.

Actions such as paying, sending, deleting, or publishing follow the selected permission rules. Full autonomy can bypass confirmation for requests from the desktop app. Review these settings before assigning work that changes or shares information.

Understand AI processing and provider costs

While this feature runs, page content and, when used, screenshots are processed by the configured model. That model can run locally or through a cloud provider. AI costs follow your provider and configuration; a separate Chrome profile does not make AI processing automatically local.

• Visit the official Codi Agent pageCheck app information, downloads, and available releases.https://agentdl.codiosity.com/
• Codi Agent version and release-date sourceOfficial release metadata supporting the version information in this guide.https://agentdl.codiosity.com/latest.json
• Meet Codi Agent, the desktop work assistantAn updated product introduction covering optional AI and skills.https://codiosity.com/en/about/insights/codi-agent-aplikasi-to-do-desktop-gratis-lokal-first