DatattentionA Jupyter server on the same host as the models
Port 8888 is the notebook server. The models are the other ports on this address.
A notebook is where you try a call, not where you serve the model. On this machine the notebook-style server is port 8888. It does not ask for a token. GET /api returns a version, GET /api/status returns the idle status, and GET / is a short HTML page.
The models are not inside that process. From a notebook on this host, or from your own machine, call Ollama on port 11434 or the OpenAI-compatible server on port 8000. Use the same URLs you would use from a shell.
Treat the notebook port as its own service. A client that only wants a completion does not need it. A person who wants the notebook UI uses http://datattention.com:8888/.
On this host
No token on port 8888. The model call beside it is still port 11434 or 8000.
curl http://datattention.com:8888/api
curl http://datattention.com:8888/api/status
curl http://datattention.com:11434/api/tagsThe full list of ports and paths is on the endpoints page.