Can't make it work: "backend offline". Any ideas?

[AlexeiMezin]

Git-cloned yesterday. Virtual machine + Ubuntu 24.04. Followed all step in install manual (with or without docker) without any errors. Local LLAMA is online with no complaints. But when I try to use web interface, I got "Backend offline • Start backend server to enable chat".

This is an example of docker start:

$ docker-compose up
Starting rag-api ... done
Starting rag-backend ... done
Starting rag-frontend ... done
Attaching to rag-api, rag-backend, rag-frontend
rag-api     | 2025-11-01 09:01:34,973 | INFO     | rag_system.indexing.contextualizer | Initialized ContextualEnricher with Ollama model 'qwen3:0.6b' (batch_size=10).
rag-api     | 127.0.0.1 - - [01/Nov/2025 09:01:37] "GET /models HTTP/1.1" 200 -
rag-backend | /app/backend/server.py:4: DeprecationWarning: 'cgi' is deprecated and slated for removal in Python 3.13
rag-backend |   import cgi
rag-frontend |
rag-frontend | > multimodal_rag@0.1.0 start
rag-frontend | > next start
rag-frontend |
rag-frontend |    ▲ Next.js 15.3.3
rag-frontend |    - Local:        http://localhost:3000
rag-frontend |    - Network:      http://172.18.0.4:3000
rag-frontend |
rag-frontend |  ✓ Starting...
rag-frontend |  ✓ Ready in 855ms

I tried direct install within python env. Everything is green, all healthchecks passed. Same result in web interface.

I connect to the web service from other machine, not from localhost. Does it matter? I use address of host machine (10.17.....:3000) and web interface is available. This is strange that docker output suggest using 172.18.0.4 which is internal docker network unavailable even from host machine.