Configuration¶
This page lists what the Vane listing tells you to set after launch. The
import steps for the model file are in /opt/sushi/docs/README.md on the
instance. This page does not add flags that the listing does not name.
Where things live¶
| Path | What it is |
|---|---|
| Port 3000 | Vane API. Reach it through Session Manager port forwarding. |
/opt/sushi/docs/demo.sh |
Walkthrough. |
/opt/sushi/docs/README.md |
Image notes, including how to import the model file. |
Mode 0600 file you create |
OpenAI-compatible chat and embedding endpoint. |
| Per-instance access file | Created for this instance. The listing does not name its path in the usage text. Read it from the image notes. |
Models¶
No model and no key are bundled. After launch, store the chat and embedding
endpoint in a mode 0600 file and import it with the guide on the instance.
Keep the key out of user data and logs.
Listen address¶
No inbound application port is required. Forward port 3000 and leave that session open. See Open port 3000.
Last updated November 1, 2026 · Suggest an improvement