Desktop quick start

Connect your first model#

1. Install and launch the Windows x64 desktop app. The packaged app includes its Python runtime; a source checkout has a separate setup step.
2. Open Settings → System Configuration to inspect the active configuration paths and add a cloud credential, or start a local model server.
3. Open Models & Reasoning and choose the chat provider and a model available to your account or local server.
4. Send a short greeting first. Confirm a response before adding document, voice, or tool workflows.

Try a small Agent task#

1. Select Agent in the chat-mode selector.
2. Enable the relevant built-in tools under Plugins & Tools.
3. Ask for a small output, such as a CSV containing three example rows and a short explanation.
4. Review an approval request if your security settings require one.
5. Open Artifacts & Workspace and confirm the file exists and contains the expected data.

Create a CSV called weekly-plan.csv with columns day and task, containing three example rows. Save it as an artifact and tell me where to find it.

For a proposal without implementation, switch to Plan before sending the next message.

Add capabilities as needed#

• Knowledge Base: ingest a small text document and ask a question whose answer is in it.
• Local voice: enable the speech plugins and download their model assets.
• Personal memory: review Auto-Learn Memories and the current saved facts.
• Connectors: install the package, start its companion where required, test the connection, and enable it.