About Rill browser
Overview
Rill is a free browser for macOS that runs Claude Code and Codex agents beside your browsing. Press ⌘E on any page to send a task or ask a question, then keep scrolling while the agent gets it done.
Rill runs on the Claude Code or Codex plan you already have, with no new subscription and no API keys. It also works as a plain browser without an agent and brings your sign-ins over from Chrome, Arc, or Safari.
Key Benefits
- Press ⌘E on any page to send a task or ask a question to your Claude Code or Codex agent.
- Rill finds the right project from your agent history, so there is nothing to set up.
- Ask about a page and the answer opens beside it from the agent you already use.
- Compare up to six tabs and get a table that quotes each page.
- Agents in a project read the web in tabs of their own, without your logins.
- Rill keeps history as files on your Mac and passwords in your Keychain, unlocked with Touch ID.
How It Works
Download Rill and drag it to Applications, then sign in to Claude Code or Codex — use either, or both. Press ⌘E on any page and Rill finds the right project from your agent history, so each request lands where the work goes while you keep browsing.
Use Cases
- Developers send code tasks from any page and keep browsing while the agent works.
- Teams route requests into the right project automatically using Claude Code and Codex history.
- Anyone on a page asks about it and reads the answer beside the page.
- Reviewers compare up to six tabs in one quoted table.
- People ask for non-code work, like a chart on their data or a flight booking, and the agent handles it in a tab behind theirs.
Why Choose This Product
Rill runs your own Claude Code or Codex in your own projects, with agent features requiring only Claude Code 2.1.261 or later, or the Codex CLI 0.144.4 or later. New tasks are read-only until you allow edits, and agents on the web stop before any step that pays, places an order, sends, posts, or deletes unless you plainly told them to go through with it.
Rill browser Pros & Cons
- Free, with no new subscription or API keys
- Uses the Claude Code or Codex plan you already have
- Automatically finds the right project from agent history
- New agent tasks are read-only until you allow edits
- History and passwords stay on your Mac
- AI features require Claude Code or Codex
- Requires Apple silicon and macOS 14 or later
- Prompt injection risk is lowered, not removed
Key Features
Task from any page
Press ⌘E on any page to send a task or question to your Claude Code or Codex agent.
Auto project detection
Rill finds the right project from your Claude Code and Codex history, with nothing to set up.
Ask about a page
Ask a question about the page you are on and the answer opens beside it from your agent.
Compare tabs
Pick up to six tabs and get a table that quotes each page.
Web-browsing agents
Agents in a project read the web in tabs of their own, without your logins, and you can take one over.
Journal-like history
Your history reads like a journal of where you went, what you asked, and what got done.
Project map
Every project your agents have touched appears on one map, with finished work marked Done.
On-device privacy
History stays as files on your Mac and passwords live in your Keychain, unlocked with Touch ID.
Rill browser Pricing
Pricing extracted from the product website and may change. Check the source for current details.
Frequently asked questions about Rill browser
What does Rill cost?
Rill is free. AI features run on the Claude Code or Codex plan you already have, so there is no new subscription and no API keys to set up.
Do I need Claude Code or Codex to use Rill?
For the AI features, yes. Browsing works without them, so you can use Rill as a plain browser even if you have no agent plan.
Will an agent change my code without asking?
New tasks are read-only until you allow edits, and in git repositories each edit can be undone.
Can an agent buy something by itself?
A task on the web is told to stop before the step that pays, places an order, sends, posts, or deletes, unless you plainly told it to go through with it. It cannot type into password, card, or one-time-code fields.
What about prompt injection?
Agents read pages as untrusted text. In a project, an agent browses in tabs of its own without your logins and clicks or types only when you allow it. A task on the web works with your sign-ins within those limits. That lowers the risk; it does not remove it.
How do I get updates?
Rill lets you know when a new version is out.