Skip to main content

Getting Started

This page covers the first run of the AI Assistant: opening it, completing the setup wizard, and asking your first question.

Open the assistant​

  • Click the AI Assistant icon in the title bar (top-right of the window), or
  • Press Ctrl + Alt + A.

The assistant opens as a panel on the right side of the window. You can move it, dock it on the left, or float it like any other Volvicon panel.

Complete the setup wizard​

The first time you open the assistant, a short wizard asks three questions:

StepQuestionOptions
1Where documentation search runsLocal (on this computer) or Cloud
2Which AI answers your questionsOllama (local), Anthropic Claude, or OpenAI / compatible
3Your model or keyA local model to use, or an API key for a cloud service

When you finish the wizard, the assistant starts and a status indicator turns green when it is ready.

Recommended for most users

Choose Local documentation search and Ollama (local) AI when your computer has a dedicated graphics card (8 GB of memory or more is a good starting point) and at least 16 GB of system RAM. This keeps everything on your computer, requires no API key, and works offline. Local AI still runs on lighter hardware, just more slowly. See Local AI.

You can change any of these settings later from the assistant's Settings tab.

Ask your first question​

Type a question and press Enter. For example:

  • "How do I export a surface to STL?"
  • "What is the difference between a mask and a volume?"
  • "Write a script to create a sphere primitive."

The first answer after opening the panel — or after switching models — takes a few extra seconds while the AI model loads into memory. This is expected. Later answers in the same session are faster.

Formatting in the chat​

The chat panel formats messages to make code and structure easy to read:

  • Code blocks. Python scripts are shown with syntax highlighting in a bordered block with a copy button in the top-right corner. Click it to copy the script to the clipboard.
  • Inline code. Wrap text in single backticks to show it as inline monospace code — for example, `get_mask_uint8`. This works in your own messages, so you can paste code or refer to API names clearly.
  • Code in your messages. You can paste a script wrapped in triple backticks, and it is shown with the same highlighting and copy button as the assistant's scripts.
  • Lists. Bulleted and numbered lists in a reply are shown as lists.
  • Emphasis. Key terms may appear in bold, with occasional italics for light emphasis.

To adjust the text size, hold Ctrl and scroll the mouse wheel, or press Ctrl + and Ctrl -. Ctrl + 0 resets it to the default. This works from either the conversation area or the message box below it, and both always use the same size. Volvicon remembers your chosen size and restores it the next time you open the assistant.

The assistant's text size is independent of Volvicon's application font size (set in File → Preferences → User Interface). Changing one does not affect the other, so you can enlarge the chat for comfortable reading without resizing the rest of the interface, and vice versa.

Chat history​

Your conversations are saved automatically on your computer. On the Chat tab, use the History panel on the left to reopen a past chat, and the New chat button (bottom-right) to start a fresh one. Right-click any entry in the History panel to rename or delete it.

Chat history is stored in a single file on your computer:

%APPDATA%\Volvicon\Volvicon\AIAssistant

Paste that path into the File Explorer address bar to open the folder. The conversations are stored there in chat-history.json.

  • Back up: with Volvicon closed, copy chat-history.json to a safe location (for example an external drive or a backup folder).
  • Restore: with Volvicon closed, copy your saved chat-history.json back into the same folder, replacing the file that is there. This also lets you move your chat history to another computer.

Closing Volvicon before you copy ensures you capture the most recent conversation. Deleting the file removes all saved chats.

The history file itself is not uploaded. If you use a cloud provider and continue a saved conversation, recent conversation text is sent to that provider as context for the next answer. See Privacy & Knowledge Retrieval.

Hardware notice​

When you choose local AI, Volvicon checks your graphics hardware and system memory. On supported graphics cards with enough dedicated memory, it uses the GPU automatically; otherwise local models run on the CPU, often taking tens of seconds per answer. If no suitable graphics card is detected and your computer has less than 16 GB of system RAM, it shows a notice that local AI may be slow. You can dismiss the notice and continue, but a cloud provider is usually a better experience on integrated-graphics or lower-memory machines.

Two independent choices

Documentation search (Local or Cloud) and the AI provider (Local or Cloud) are separate settings. For example, you can keep documentation search local while using a cloud AI, or the reverse. See Privacy & Knowledge Retrieval.

Next steps​