This project provides a browser-native side panel extension for Hermes Agent, enabling users to seamlessly connect active web context to their local or remote Hermes runtime. It targets users of the Hermes AI framework who need to integrate browser data into their AI workflows, offering a direct bridge for context-aware AI interactions.

How It Works

The extension functions as a Chrome/Edge MV3 side panel, leveraging the Side Panel API to connect to a configurable local or remote Hermes Agent API server. It synchronizes Hermes providers, models, and sessions, then captures detailed context from the active browser tab—including URL, title, selected text, and page structure—to feed into the Hermes runtime. It supports streaming responses and offers various UI themes and quick commands for common web-based AI tasks.

Quick Start & Requirements

Installation involves cloning the repository, running `npm install`, and building the extension with `npm run build`. The resulting `dist/` folder is then loaded unpacked into Chromium-based browsers (Chrome 114+ baseline). Prerequisites include a running Hermes Agent with its API server enabled and Node.js 20+. Official documentation links are available for Hermes Agent and its API server.

Highlighted Details

• Chrome/Edge MV3 side panel architecture.
• Connects via REST API (local/remote) or WebSocket (remote dashboard).
• Synchronizes Hermes providers, models, profiles, skills, and sessions.
• Captures comprehensive active tab context: title, URL, selected text, page DOM elements (links, forms, buttons), and metadata.
• Provides quick commands: /summarize, /explain, /rewrite, /tabs, /action-items.
• Supports voice dictation via Hermes STT or browser speech fallback.
• Offers desktop-style appearance settings with multiple themes.
• Employs a conservative security model with read-only context capture and minimal browser permissions.

Maintenance & Community

This is a community extension created by Jon Komet (@abundantbeing) for Hermes Agent by Nous Research. It is currently in public alpha status.

Licensing & Compatibility

The project is licensed under the MIT License, permitting commercial use and integration with closed-source applications.

Limitations & Caveats

This project is in public alpha and not yet available on the Chrome Web Store. Remote dashboard WebSocket connections have limitations, excluding image attachments and skills/profile listings. Browser internal pages and sensitive categories are intentionally restricted from context capture. The extension explicitly avoids permissions for click/type actions, form submission, cookie access, history, downloads, or bookmarks, and does not offer browser control capabilities.