Follow the steps in order. Each has one command and the result to expect.
python3 -m venv nbx && source nbx/bin/activate
pip install neurobrix
neurobrix doctor
Expected: No blocking problem found. On a GPU machine, read Install by hardware first. Problems: Troubleshooting.
neurobrix hub --category UPSCALER
Expected: a list of upscalers with size and licence. The same rows are on neurobrix.es/catalog. The categories are listed there too; an unknown one is refused with the list of valid ones.
neurobrix import caidas/swin2SR-classical-sr-x4-64
Expected: an IMPORT COMPLETE block. import takes the full slug org/Model, as the hub spells it. run, serve and upscale take the bare name (here swin2SR-classical-sr-x4-64).
Check disk first if space is tight: an import needs about twice the archive size plus 2 GiB.
neurobrix upscale --model swin2SR-classical-sr-x4-64 --input photo.png --output photo-x4.png
Expected: photo-x4.png, four times larger on each side. No GPU flags are needed.
neurobrix import TinyLlama/TinyLlama-1.1B-Chat-v1.0
neurobrix serve --model TinyLlama-1.1B-Chat-v1.0
neurobrix chat
neurobrix stop
Expected: serve starts a background daemon that loads the model once; chat talks to it (/quit leaves); stop frees the memory. See Serving.
neurobrix hub --category IMAGE
Pick one that fits your VRAM, import it, then:
neurobrix run --model <Model> --prompt "A lighthouse at dusk" --steps 20 --seed 42 --output lighthouse.png
Expected: lighthouse.png. See Inference for the parameters.
neurobrix list
neurobrix remove <Model>
Expected: the model leaves ~/.neurobrix/cache/; the archive stays in store/ unless you also use neurobrix clean --store.
Run your first model · Model hub · CLI reference · Licences
© 2025–2026 Hocine Benkelaya. NeuroBrix engine released under the Apache License 2.0.
Developed by Hocine Benkelaya · Supported and funded by Neural Network Holding Ltd · Commercial distribution by WizWorks OÜ