0xSero/parchi

What it solves

Parchi is an AI-powered browser copilot that allows users to automate browser tasks through a chat interface. It eliminates the need for manual navigation and interaction by enabling the AI to perform actions like clicking, typing, and extracting information directly within the browser.

How it works

It operates as a Chrome or Firefox extension that integrates a side panel chat. Users configure an OpenAI-compatible model provider (such as OpenAI, OpenRouter, or local models via Ollama) to power the AI. The system uses an orchestrator and subagent flow to execute browser tools—including navigation, reading, interacting, and taking screenshots—to accomplish the user's requests.

Who it’s for

Users who want to automate repetitive web-based tasks, extract data from websites, or summarize web content using an AI agent that can interact with the browser in real-time.

Highlights

• Browser Automation: Ability to navigate, read, click, type, and extract data.
• Flexible Model Configuration: Supports any OpenAI-compatible endpoint, including local LLMs.
• Multi-browser Support: Compatible with Chrome (MV3) and Firefox.
• Control and Safety: Includes tool permissions and domain allowlist controls to manage AI actions.