OLLAMA / ON YOUR COMPUTER
Forge builds, Pixel refines and Atlas explains. One model installed on your computer powers all three, without a paid API key or a FuturAI account.
AI requests go directly to Ollama on your computer. Results stay in this tab and are not published to the site. Download them before leaving or reloading the page. Speed and quality depend on your hardware and model.
On a Windows, macOS or Linux computer. This workshop uses the model on the computer where the site is open.
ollama pull qwen2.5-coder:3bIf you already have a local model, keep it and continue. Otherwise, about 1.9 GB to download, plus memory to run it. No API fees; your computer provides the compute. A larger coding model may improve results if your machine supports it. Model details
$env:OLLAMA_HOST='127.0.0.1:11434'
$env:OLLAMA_ORIGINS='https://futurai.space'
$env:OLLAMA_NO_CLOUD='1'
ollama serveKeep this window open while using the agents. Run this command again for each new session; the model stays installed. If the port is already in use, quit the other Ollama instance first. Ollama help
Ollama not connected
An adjustable focus timer with start, pause, reset and a guided break. No personal tracking.