ADOS

Quick reference for connecting to and using a brev.nvidia.com GPU instance.

Key Paths

Path	Description
/models	Model checkpoints
/ComfyUI	ComfyUI installation
/LTX-2	LTX-2 installation
/mnt/gcs	GCS bucket mount
/scripts/sync_models.sh	Manual model sync script

---

Access

Web Terminal (ttyd)

Port: 7681

https://web-ssh0-<instance-id>.brevlab.com

ComfyUI

Port: 8188

https://comfyui0-<instance-id>.brevlab.com

To start:

cd /ComfyUI
source .venv/bin/activate
python main.py --port 8188 --disable-cuda-malloc --listen 0.0.0.0

JupyterLab

Use for uploading/downloading files or running notebooks.

Port: 8889

https://jupyter0-<instance-id>.brevlab.com

Upload:

Download:

---

GCS Bucket

The bucket is auto-mounted at /mnt/gcs on container startup

To sync models manually:

/scripts/sync_models.sh