Your first chat

The first time Froots opens, it runs a short setup inside the main window. Here's what each step does, and what happens right after.

The setup

1. Welcome — "Your people, your agents, and your work — one place, running on your computer, owned by you." Click Get started and Froots creates your identity key on this machine. Already have a key from another computer? Choose Use an existing key instead and paste it.
2. Back up your key — the key is your account. There's no email, no password, and no copy on a server, so saving it is the only way to get the account back. You can skip this and do it later in Settings → Profile, but don't put it off.
3. Choose your default harness — pick the agent CLI your agents run on: Claude Code, Codex, Gemini CLI, and any others on your machine. A green dot means it's installed. If it isn't, the step shows how to set it up and a Check again button. Details in Agent runtimes.
4. Your profile — your name and picture. This is how people and agents see you.
5. Meet your agent — your first agent gets a random name. Keep it, roll a new one, or type your own, then pick what it should help with: Assistant, Writer, Researcher, Coder, or Productivity.
6. Connect your accounts — optionally connect Google (mail and calendar) and GitHub (pull request review and pushing from the Dev tab). Both are skippable and live in Settings → Accounts.
7. Default browser (macOS) — optionally make Froots your default browser, so links open inside Froots where your agents can see them. Skippable.
8. Done — "Your agent is ready." Click Start using Froots.

Starting the conversation

After setup you land in the main window. The sidebar is your Inbox: anything waiting on you, with a dot for anything unread. Swipe sideways for Pinned. A few ways in:

• Click your agent's face in the sidebar, or press ⌘J to open your first agent.
• Press ⌘K to find anything.
• Press the global hotkey, ⌘⇧Space by default, from anywhere. It brings Froots forward with the chat focused.

Ask for something small first: summarize this folder, draft a note about X, check what's on my calendar tomorrow. Watching a small task end to end is the fastest way to get a feel for what the agent can do.

When the agent offers a suggested reply, it shows as faint text in the message box. Tab fills it in, Enter sends it. You can reply to, forward, or react to any message.

Friends and groups

People and agents share the same chats. To add a friend, open Profiles, click Add friend, and paste their key (it starts with npub). They get a friend request, and you're friends once they accept. Your own key is in the same menu, so you can send it to someone.

To start a group, select a mix of friends and agents in Profiles and start a group. Agents in a group answer when you @mention them.

What you'll see during a run

Froots shows the run as it happens: tool calls, file edits, and steps stream into the chat as the agent works. In a one-on-one chat with your agent, the picker in the top-right shows which harness and model it's on. Click it to switch. If a turn fails, the chat says why and offers Retry.

Once small tasks feel trustworthy, hand over bigger ones, or schedule them to run on their own with Routines.