Feels like a cloud app. Runs on your machine.
Everything you'd expect from a hosted assistant, without handing your data to one.
Chat from anywhere
Open a conversation with whichever model is running on your hardware, from your laptop, your phone, or someone else's computer.
Qadenza finds the fastest path to your machine automatically, so it feels like a hosted app instead of a home network project.
Models that fit, no token limits
Browse models and see whether each one fits your hardware before you download it. Install with one click, on Ollama or llama.cpp.
Got more than one GPU? Split a model across them. There's no meter running: use it as much as your hardware can handle.
Use it where you already work
The VS Code extension turns your own models into a coding agent: it reads and searches your code, proposes changes, and shows every edit as a diff for you to approve.
Qadenza also speaks MCP and the OpenAI API, so Claude Code, Cursor and anything with a custom base URL can use the models on your machine.
Your files are read and edited locally. Only the conversation goes to your model.
Let any MCP client check your servers and hand work to your local models.
Tokens are revocable any time from the dashboard.
Everything else, in one place
Pull models, watch temps and load, generate images, and see what's connected across every machine you own. Agents update themselves.
One dashboard for all of it, whether you're checking in from your desk or from across the country.
