Quick Start

Quick Start

This page takes you from a fresh install to your first answer: sign in, start a conversation, send a message, pick a model, and see the work.

Already installed FutureOS? If not, start with Install FutureOS.

---

1. Open and sign in

To use the built-in models, connect your FutureGene account:

1. Click the gear icon (Settings) at the bottom-left of the window.
2. Go to the Providers page.
3. Under Built-in, find FutureGene and click Connect.
4. Your browser opens for authorization. If it doesn't open automatically, the app shows a verification code and a copyable link — open the link in any browser and enter the code.
5. Once authorized, the app confirms you're connected.

Prefer your own model provider? You can add an OpenAI-compatible or Anthropic-compatible provider instead. See Settings.

---

2. Start a conversation

There are two ways to begin, and you pick from the same screen:

	New Chat	Workspace
Best for	Quick questions, one-off tasks	Real projects tied to a folder
Setup	Nothing — just type	Pick a folder on your computer

• Click New Chat (top-left) for the fastest start.
• To work on a real project, choose Workspace, then Open a folder on your computer. The agent works inside that folder.

---

3. Send your first message

Type in the input box at the bottom and press Enter to send. You'll see:

• The reply streaming in live.
• Tool activity shown as the agent reads files, runs commands, and so on.
• A pause for your approval whenever the agent wants to do something risky (like writing a file or running a command). Nothing risky happens without your OK.

You can attach up to 4 image files per message — use the paperclip button, paste an image, or drag files onto the input box.

---

4. Pick a model (optional)

The model selector sits right inside the input box — click it to switch models for this conversation. There's also a thinking level control next to it.

To manage which models appear, open Settings → Models. See Settings.

---

5. See the work

Open the panel on the right to check what the agent is doing. It has up to three views:

• Runs — background programs the agent is running, with the real commands shown. You can stop a running one or clear finished ones.
• Review — for a Workspace, the file changes the agent made: the file list, stats, and diffs.
• Artifacts — for a Chat, the outputs the agent produced (reports, summaries, tables, files).

---

See also

• Using FutureOS — a full tour of the app.
• Settings — providers, models, and desktop options.
• Skills — capability packs the agent can use.