A new community-built browser extension is making it possible to bring Hermes AI into every Chrome or Microsoft Edge tab without switching back and forth between browser windows.

The project, called Hermes Browser Extension, was created by developer Jon Komet and is currently available as an open-source public alpha. It's designed to work with the Hermes Agent runtime from Nous Research, adding a native browser side panel that can send page context directly to a local or remote Hermes instance.

Rather than acting as another browser chatbot, the extension works as a front end for an existing Hermes setup. Once connected to a Hermes Gateway, it can access the models, sessions, skills and tools already configured on the backend. That means users can switch between providers such as GPT, Claude Opus, Grok or even locally hosted models without leaving the browser.

In a post on X, Komet described it simply as Hermes that "lives on every tab."

browser extension for hermes agent @NousResearch

hermes lives on every tab🪽

- side panel chat on any page
- swap models: opus, gpt, grok, local
- session picker, vision + screenshots
- connects to your local or remote hermes gateway
- unofficial, open source, loved by… pic.twitter.com/6volmtIGL0

— komet💫 (@jonkomet) June 25, 2026

Built around your existing Hermes setup

The side panel can read the active page, capture selected text, collect headings, links, forms and other readable content, then package everything as browser context before sending it to Hermes. It also supports screenshots, voice dictation and session management, while allowing connections to either a machine running Hermes locally or a remote server over a trusted connection.

Although it's still early software, development is moving quickly. Komet announced version 0.1.4 shortly after the initial release, adding more reliable side panel behavior along with automatic session names, live gateway status, update checks that compare installed and available versions, and trusted scans for private hosts and Tailscale-connected agents.

The extension isn't available through the Chrome Web Store yet. Installation currently involves cloning the GitHub repository, building it with Node.js, and loading the generated dist folder as an unpacked extension through Chrome or Edge's developer mode.

Security questions come up early

Its permission model has also attracted attention. The extension captures page content so Hermes can understand what's on screen, but according to its documentation it deliberately avoids permissions for cookies, browsing history, bookmarks, downloads, native messaging or browser automation. It also refuses to collect context from browser-internal pages and blocks obvious categories such as banking, cryptocurrency, password managers, payment portals, government tax sites and health-related pages.

Still, not everyone is convinced that's enough.

One X user pointed out that an extension with access to page content and screenshots could theoretically view autofilled credentials or capture sensitive login pages if misused, especially since this remains an unofficial project. Komet didn't dispute the concern directly, though he replied elsewhere that the long-term goal is for the extension to become an official part of the Hermes ecosystem.

For now, Hermes Browser Extension sits somewhere between an experiment and a preview of what's coming next. The GitHub repository also hints that the groundwork may eventually lead to a standalone Hermes browser for macOS, Linux and Windows, though that's described as a future project rather than something currently under development.