A small, fully-customizable UI for chatting with your local Ollama models. Plain HTML/JS frontend plus a thin FastAPI proxy — no build step, no cloud, no API keys.
The studio runs on local models — Ollama everywhere — so it needs a chat client that's local-first and
endlessly tweakable rather than a locked-down product. my-AI-chat is exactly that: a couple of files, no
toolchain, and every knob exposed. Run it, open localhost:8010, and talk to whatever you've
pulled into Ollama.
/api/models, /api/chat streaming, /api/presets). First run makes a
venv and installs FastAPI/uvicorn/httpx; you just need Ollama on localhost:11434.
System prompt, temperature, top-p, max tokens — editable live, per chat.
Bundle a model + prompt + params under a name, switch with one click. Saved to data/presets.json.
Four base themes plus a color picker for every token, font size, and corner roundness. Saved in the browser.
⊞ Compare sends one prompt to several models at once, side by side.