Docs menu
Rodatacenter
Use the team's self-hosted AI chat at ai.sfuracerbot.ca from any browser, with no account.
Rodatacenter is the team’s own AI chat, running on a workstation in the lab. It is open to everyone at ai.sfuracerbot.ca, with no sign-in and no access request. The workstation has no usable GPU, so small language models run on its CPU: everyday questions get quick answers, and long or hard ones are slower.
Start a chat
- Open ai.sfuracerbot.ca.
- Pick a model from the dropdown in the message box. LFM2.5-8B-A1B, the default, suits most questions; Models compares the four.
- Send a message. The first reply after a model switch waits 10 to 40 seconds while that model loads.
Chats are saved in your own browser only, so another device, or clearing site data, starts with none. Settings → Import/Export moves chats between devices.
Sharing the machine
The server loads one model at a time and runs two chats at once. Switching models makes everyone else’s next reply wait for the switch too, so stay on the default model when several people are using it.
- ModelsPick the right one of Rodatacenter's four models for a question, and set how long it thinks before answering.
- APICall Rodatacenter from code through its OpenAI-compatible API, with streaming and the per-request controls it accepts.
- How it worksThe server behind the chat, why a CPU-only workstation can run these models, and what each error means.