AI Assistant

MotioForge Studio includes a conversational assistant that answers questions about using the software, hardware configuration, and troubleshooting. The assistant knows the entire documentation and can automatically read the current session log to give contextual answers to errors.

How to open the assistant

Press F2 from any MotioForge Studio window.
Click the Ask AI button inside the help window (F1).
From the menu Help → Ask AI.

What it can answer

How to configure a domain, port, or element.
How to create and edit animations on the timeline.
What an error or warning message means.
How to solve a connection or playback problem.
Ideas and suggestions for possible applications of the system.

Note: The assistant only answers questions relevant to MotioForge Studio and animatronics. It does not assist with programming, source code, or topics unrelated to the system.

Available backends

The assistant supports two AI engines, selectable in Preferences → AI Assistant:

Backend	Cost	Requirements	Privacy
Ollama (local)	Free	Ollama installed on the PC + model downloaded (~2–4 GB)	No data leaves your local network
Claude API (Anthropic)	Paid by the user (fractions of a cent per session)	Anthropic account and personal API key	Data sent to Anthropic servers

Note: if a key is entered in the Claude API key field, MotioForge Studio uses Claude automatically, regardless of the selected backend.

Setting up Ollama (free, local)

Ollama is an open-source program that runs AI models directly on your computer, with no Internet connection and no cost.

1 — Installation

Go to https://ollama.com and click Download.
Choose the package for your operating system:
- Windows — download OllamaSetup.exe and run the installer. After installation Ollama is available in the Start menu and the system tray.
- macOS — download the .zip file, extract the app and drag it to Applications. On first launch it may ask for permission to run third-party software.
- Linux — open a terminal and paste:
  curl -fsSL https://ollama.com/install.sh | sh

Hardware requirements: Ollama works on any modern PC with at least 8 GB of free RAM. A dedicated Nvidia or AMD GPU delivers responses in a few seconds. Without a dedicated GPU (CPU-only), each response can take 1 to 5 minutes, even on powerful machines. If wait times of this length are a problem, use Claude API (section below): responses arrive in 5–15 seconds regardless of hardware, at negligible cost.

2 — Downloading a model

The "Launch" window that appears after installation is not needed for MotioForge Studio. It shows standalone Ollama apps (such as Claude Code, Codex, Hermes…) that run in the browser. You can close it and continue with the step below.

Open a terminal (Command Prompt on Windows, Terminal on macOS/Linux) and download a model with one of the following commands:

Command	Size	When to use
`ollama pull llama3.2`	~2 GB	Good balance between quality and speed. Recommended to start with.
`ollama pull mistral`	~4 GB	More detailed and accurate answers.
`ollama pull phi3`	~2 GB	Very lightweight, suitable for PCs with limited RAM.

The download takes a few minutes depending on your connection. Once complete, the model is stored locally and does not need to be downloaded again.

3 — Starting Ollama

Ollama must be running before using the MotioForge Studio assistant.

Windows: Ollama starts automatically with Windows after installation. If it is not running, search for Ollama in the Start menu and launch it. An icon will appear in the system tray (next to the clock).
macOS: Open Ollama from the Applications folder. The icon will appear in the menu bar.
Linux: In a terminal type: ollama serve
Leave the terminal open or start Ollama as a system service.

To verify that Ollama is active, open a browser and go to http://localhost:11434. If you see the text Ollama is running, everything is working.

4 — Configuration in MotioForge Studio

Open Preferences → AI Assistant.
Select Ollama (local / free) as the backend.
In Ollama URL leave the default value http://localhost:11434 (change it only if Ollama runs on another computer on the network).
In Ollama model choose or type the name of the downloaded model, e.g. llama3.2.
Click OK and press F2 to open the assistant.

Setting up Claude API (higher quality answers)

Claude is the AI model developed by Anthropic. It provides significantly more accurate and contextual answers than local models, especially when analysing error logs. The service is paid, but the cost for typical usage is negligible.

1 — Create an Anthropic account

Go to https://console.anthropic.com and sign up.
In the Billing section, add a payment method and top up your credit (even €5–10 is enough for months of normal use).

2 — Generate an API key

In the Anthropic console open the API Keys section.
Click Create Key, give the key a name (e.g. MotioForge) and copy the value.
Store the key in a safe place: it will not be visible again after closing the page.

3 — Configuration in MotioForge Studio

Open Preferences → AI Assistant.
Select Claude API (Anthropic) as the backend.
Paste the API key into the Claude API key field.
Choose the model:
- Claude Haiku — economical and fast (cost ~€0.001 per question). Recommended for everyday use.
- Claude Sonnet — higher quality, deeper log analysis (cost ~€0.01 per question).
Click OK and press F2 to open the assistant.

A session of 10 questions with Claude Haiku costs less than €0.01. With Claude Sonnet it stays under €0.10. The cost depends on the length of answers and the number of log lines included in the context.

Automatic log reading

When the Log lines included field in preferences is greater than 0, MotioForge Studio automatically attaches the last lines of the log file to the context sent to the AI. This enables contextual answers to real session errors:

"Why doesn't the port connect?" — the AI reads the exact error message.
"What does this warning mean?" — the AI interprets the log text.
"Playback always stops at the same point, why?" — the AI analyses the event sequence.

The number of lines included is configurable in preferences (default: 80 lines; 0 = disabled).

Privacy and security

Ollama: no data is sent to external servers. Everything stays on the user's PC.
Claude API: the question text, documentation, and log lines are sent to Anthropic's servers. Do not include personal data or confidential information in your questions.
API key: the key is stored in the user preferences file. Do not share that file with others.