Open source · MIT License · Free forever

Your browser,
on autopilot.

Chat naturally in the side panel — doeverything navigates, clicks, fills forms, and executes tasks across every tab. Works with every major AI provider.

Add to Chrome — It's free View on GitHub

New issue · github.com

github.com/acme/repo/issues/new

Doe

New issue

Title *

Label

Assignee

Unassigned

Description

doeverything

gemini-2.5-flash

Ask Doe to do anything…

What you can do

Automate anything.
Your way.

From one-off tasks to daily workflows — describe what you want, watch it happen.

Deep Research

Ask the AI to research any topic. It opens tabs, reads sources, cross-references data, and delivers a synthesis — without you lifting a finger.

"Find me the 5 best noise-cancelling headphones under $300 and compare their specs"

Form Autopilot

Fill any form on the web with one instruction. Applications, registrations, bug reports — the AI navigates and fills every field for you.

"Submit a support ticket about the login issue with these details: …"

Workflow Skills

Save any multi-step workflow as a reusable Skill. Activate it with a slash command — saves hours on tasks you repeat every day.

/daily-standup → posts to Slack, checks Jira board, opens today's calendar

Email & Calendar

Draft and send emails, RSVP to invites, schedule meetings — describe the intent, the AI handles all the clicking, typing, and tab navigation.

"Reply to Sarah's meeting invite and suggest Thursday at 3 PM instead"

Smart Shopping

Track prices across retailers, find coupon codes, compare products side by side. Stop switching tabs — the AI brings the comparison to you.

"Find the cheapest iPhone 16 Pro 256GB across Amazon, Best Buy, and Apple Store"

Dev Assistance

Inspect console errors, analyze network requests, execute JS snippets, and capture screenshots. Your AI pair programmer that uses the actual browser.

"Find what's causing the 401 on that API call and show me the request headers"

Simple by design

Three steps to
browser automation

Install & connect your AI

Add the extension from the Chrome Web Store. Enter your API key for any supported provider — Claude, GPT-4, Gemini, Groq, Mistral, and more. Your keys never leave your device.

Open the panel, describe the task

Click the extension icon to open the side panel on any page. Type what you want done in plain language — no special syntax, no code required.

Watch it happen

The AI navigates, clicks, types, reads, scrolls, and reports back. You can interrupt, redirect, or hand over at any point — you're always in control.

Reusable automations

Save workflows as
Skills

Write a skill in Markdown once. Activate it forever with a slash command.

My Skills

daily-standup

deep-research

send-email

fill-form

price-tracker

---
name: Daily Standup
description: Post my daily standup to Slack
---

Navigate to our Slack workspace and post my standup
update in the #engineering channel:

**Yesterday:** {{yesterday}}
**Today:** {{today}}
**Blockers:** {{blockers or "None"}}

Then open Linear, filter Issues by "In Progress"
assigned to me, and screenshot the board.
Finally, open my Google Calendar and list
today's meetings.

MCP mode

Use with any
AI client

doeverything exposes all 35 browser tools as an MCP server — so any AI client can navigate, click, and read your browser while you work.

Cursor, VS Code, Claude Code, ChatGPT, Windsurf, Zed
All 35 browser tools available over HTTP or stdio
Secure cloud relay — no port forwarding needed

{
  "mcpServers": {
    "doeverything": {
      "command": "npx",
      "args": [
        "-y",
        "doeverything-mcp"
      ]
    }
  }
}

Your browser,on autopilot.