What are we working on?
Ask a question, attach a log or code file, or switch models to get a second opinion in the same conversation.
Attach up to six UTF-8 text, code, or log files, up to 96 KB each. Messages and attachments are sent to the Space to run OpenHarness.
Settings
Model access
The token stays in this page. Bimo uses it through the Space, and XORTRON sends it directly to Hugging Face for the download.
XORTRON WebGPU
Open the standalone Space for WebGPU access.
GPU memory settings
Choose fewer layers to share work between the GPU and CPU.
The default Q3_K_M file is about 13.5 GB.
Chat memory
Both models share recent chat history and attached files. This browser saves the conversation and files when storage is available. OpenHarness can compact long context during a turn.