Note: this address points to the webmcu-ai GitHub organization. If you are hosting this page from your own GitHub account or organization instead, replace https://webmcu-ai.github.io with your own GitHub Pages address before running the command.
3
Windows only: reboot your computer. The environment variable does not take effect until after a full restart. (Mac/Linux: the export command takes effect in that terminal session right away — no reboot needed.)
4
After rebooting, wait for Ollama to start and show its icon in the system tray (Windows) or menu bar (Mac). Then load the installed PWA. Return here — the status pill top-right will turn green when connected.
5
Pull a model using the field below, or run this in a terminal:
ollama pull gemma4:12b
Note: Ollama model tags use size shorthand — gemma4:e2b, gemma4:12b.
Preparing...
▶System Prompt(optional — sent as first message every conversation)
History: 0 turns
· Active: —
· ctx: — tokens
conversation
◎
No messages yet
Verify Ollama is running, then type or attach files below
or drag & drop files anywhere
⚙Output Tools
▼ show
Send last response to another AI (copies text, opens tab)
Tip: select specific text in the chat first to send only that; otherwise the full last response is copied.
HTML Extractor
Math Renderer
Image / SVG
MathJax rendered output will appear here...
SVG or base64 image from the last response will appear here...