RIRD Docs

Documentation

How to use RIRD, the autonomous AI browser agent.

Browser Agent

rird.ai

The AI browser agent is embedded directly on the homepage. Open a session, type what you want, and the agent browses, clicks, fills forms, and reports back.

1
Session starts automatically

A live Chrome session launches when you load the page. The browser panel shows the live screen. If it is blank, wait a few seconds or click "Launch Browser".

2
Type a task in the chat

Use plain English. "Go to amazon.com and find the best-rated coffee grinder under $50" or "Search LinkedIn for CTOs at fintech startups in NYC". Be specific for best results.

3
Watch it work, then read the result

The agent shows each step it takes. When done it sends a short summary in the chat. You can keep the conversation going or give a new task.

Window controls
  • Red dot -- close the session and start fresh.
  • Yellow dot -- hide/show the chat panel so the browser fills the whole window.
  • Green dot -- toggle fullscreen.
  • Click the URL bar to type a URL and press Enter to navigate manually.
  • Tabs are supported -- the agent can open, switch, and close tabs.
Mobile (iPhone / Android)
  • Trackpad cursor -- drag your finger on the browser panel to move the remote PC cursor. Tap to click. Two-finger tap for right-click.
  • The chat panel scrolls normally. Tap any message to copy it, or use "Copy chat" to grab the full conversation.
  • The layout fills your entire screen -- browser on top, chat on bottom.
Skills & Slash Commands

Built-in skills

Type / in the chat input to open the skills menu. Built-in skills give the agent a pre-filled prompt you can edit before sending.

/search
Search the web for a topic and return top results with key details.
/navigate
Go to a URL and report what is on the page.
/research
Research a topic across multiple sources and return a full summary.
/summarize
Take a snapshot of the current page and summarize the key content.
/find
Find something specific on the current page.

Save any prompt as a custom skill using the bookmark icon in the input bar. Your skills are stored locally and appear in the / menu.

Automation

Recurring tasks

The clock icon in the input bar opens the recurring task panel. Set a prompt and an interval (in minutes) and the agent will run it automatically on a loop for as long as your session is open.

Examples
  • Check the price of BTC every 5 minutes and alert me if it drops below $X.
  • Every 30 minutes, go to my competitor's site and report any new listings.
  • Every 60 minutes, search Twitter for mentions of my brand and summarize.
Email

Sending email

The agent can send email in two ways -- from its own address or from yours.

From agent@rird.ai
The agent has its own @rird.ai email address. Just say "send an email to X about Y" and it will send immediately -- no login needed.
From your email
Save your email login (Gmail, Outlook, Yahoo, etc.) using the logins panel. The agent will log in via the browser and send from your address on your behalf.
Saved Logins

Logging into sites

The lock icon in the chat input bar opens the saved logins panel. Save credentials for any site and the agent can log in automatically when you ask it to.

1
Add a login

Click the lock icon, fill in the label, URL, login type, and username/password. Supports email, Google, Twitter/X, Facebook, GitHub, LinkedIn, and more.

2
Tell the agent to use it

Say "log into my Gmail using my saved credentials" or click "Use" next to the saved login. The agent navigates and fills in the credentials automatically.

Storage
  • Signed-in users: logins are saved to your account and persist across devices.
  • Guests: logins are saved to your browser's local storage for this device only.
Memory

Persistent memory

The agent can remember facts across sessions. Tell it to remember something and it will recall it the next time you open a session.

Signed-in users
Memories are tied to your account and load automatically at the start of every session. The agent gets smarter the more you use it.
Guests
Memories last for the duration of the browser session and are cleared when you close the tab.
Examples
  • "Remember that my timezone is PST and I prefer metric units."
  • "Save my company name and website URL so you don't have to ask every time."
  • "Remember that I'm tracking AAPL and want daily price checks at market open."
Account

/success

Your account dashboard. Shows your plan, usage, API key, and downloads.

API Key
Your rird_ license key. Used to connect the CLI tool or the desktop app.
Usage
Message credits reset weekly based on your plan.
Downloads
Desktop app and CLI installers are linked from your dashboard.
Desktop Agent

rird Desktop beta

A native desktop app powered by Tauri. No Python required. Connects directly to RIRD cloud via WebSocket for real-time AI control of your browser.

macOS (Universal)
rird_0.1.0_universal.dmg
Universal binary (Apple Silicon + Intel)
Windows (x64)
rird_0.1.0_x64_en-US.msi
Windows 10 / 11, 64-bit
Linux (x64)
rird_0.1.0_amd64.AppImage
or rird_0.1.0_amd64.deb — all releases at GitHub Releases
What is included
  • Native app (Tauri 2) -- no Python, no manual setup
  • Bundled Chromium controlled by RIRD AI
  • Real-time WebSocket protocol (wss://rird.ai/ws/agents)
  • Credential masking -- passwords never reach the LLM
  • Auto-updater built in

Legacy Desktop Agent v3.x Archived

Archived. Use the Tauri Desktop app above for new installs.

The original Python + Playwright desktop app. Still works, but will not receive new features.

Mac
curl -fsSL https://rird.ai/install-desktop-mac.sh | bash
or download install-desktop-mac.sh and run it
Windows
irm https://rird.ai/install-desktop.ps1 | iex
or download install-desktop.ps1 and right-click "Run with PowerShell"
Linux
curl -fsSL https://rird.ai/install-desktop-linux.sh | bash
or download install-desktop-linux.sh and run it
What it installs
  • Python 3.12 (if not already installed)
  • Playwright + pywebview Python packages
  • Chromium browser for AI control
  • Desktop shortcut / app launcher
Pricing

Plans

2
$20
/month
3
$40
/month
4
$200
/month
Troubleshooting

Common issues

Browser shows blank
Wait 3-5 seconds for the session to initialize. If it stays blank, click the red dot to close and relaunch.
Agent is slow
Complex tasks (multi-step navigation, form filling) take 15-60 seconds. The step indicators in the chat show what it is doing.
Site blocked / CAPTCHA
Some sites detect automation. The agent will try alternative approaches automatically. For better results on sites you are already logged into, use the Desktop app instead.
Session expired
Sessions close after 5 minutes of inactivity. Refresh the page to start a new one.
Memory not loading
Make sure you are signed in. Persistent memory requires an account. Try signing out and back in.

More help at troubleshooting guide.