[Bug]: Manual scan shows no documents, while RAG indexing works (Paperless-NGX API OK, 16 docs indexed)

[funky51]

🔍 Bug Summary

Manual scan shows no documents, while RAG indexing works (Paperless-NGX API OK, 16 docs indexed)

📖 Description

Paperless-AI can successfully connect to Paperless-NGX and the RAG engine works correctly (documents are indexed, RAG chat returns relevant results).

However:

• the Manual tab shows an empty "Choose a document…" dropdown (no documents listed),

• the automatic scanner never processes any documents,

• logs repeatedly show:

  • Failed to get own user ID. Abort scanning.
  • [DEBUG] Invalid API response on page 1
  • [DEBUG] Finished fetching. Found 0 documents.

So RAG is fully functional, but the "classic" manual/auto scan pipeline sees 0 documents.

🔄 Steps to Reproduce

1. Run Paperless-NGX behind http://PAPERLESS_HOST:8899 (Docker, Synology NAS).

2. Deploy Paperless-AI via Docker/Portainer stack with:

   • PAPERLESS_API_URL=http://PAPERLESS_HOST:8899
   • PAPERLESS_API_TOKEN=<redacted> (from Paperless-NGX user profile)
   • PAPERLESS_USERNAME=<my-paperless-user>
   • AI_PROVIDER=ollama
   • OLLAMA_API_URL=http://OLLAMA_HOST:11441
   • OLLAMA_MODEL=llama3.1:latest

3. Open Paperless-AI Web UI.

4. Go to RAG Chat and click "Start indexing".

   • → RAG indexing completes successfully (N documents indexed).
   • → RAG chat can answer questions based on these documents.

5. Go to the Manual tab.

   • The dropdown "Select document → Choose a document…" is always empty.

6. Click "Scan now" in the Manual/Scan section.

   • No entries appear in History.
   • No Paperless-NGX documents get updated tags or metadata.

7. Check the Docker logs.

   • You will see Failed to get own user ID. Abort scanning. and Invalid API response on page 1 coming from the scanner/Node part, while the RAG part reports a successful index with N documents.

✅ Expected Behavior

• The Manual tab should list Paperless-NGX documents so that I can select one and run "Analyze with AI".
• The automatic scanner should be able to fetch and process documents (according to the configured tag rules) and show entries in the History tab.
• The scanner/Node side should parse the Paperless-NGX API response in the same way as the RAG backend:
  • If RAG can see N documents, the manual/auto scan should also see N documents, not 0.

❌ Actual Behavior

• RAG: works

  • RAG indexing completes successfully:
    • The logs show a full fetch from the Paperless-NGX API, with HTTP 200.
    • ChromaDB and BM25 indices are built with N documents.
  • RAG chat returns relevant answers based on those documents.

• Manual / Auto scan: does not work

  • Manual tab:
    • "Select document → Choose a document…" is always empty (0 documents).
  • Automatic scan:
    • "Scan now" does not process any document.
    • The History tab remains empty.
    • No documents in Paperless-NGX receive "AI processed" tags or updated metadata.

🏷️ Paperless-AI Version

3.0.9

📜 Docker Logs

_Excerpt (redacted and shortened):_


Fetching documents from Paperless-NGX API: http://PAPERLESS_HOST:8899
Response status code: 200

Processing documents: 16/16
Saved 16 documents to ./data/documents.json
ChromaDB collection contains 16 documents
BM25 index contains 16 documents
Search engine validation successful

Context request: Which documents are about credit?
Semantic search found 16 results
Hybrid search found 16 results
Reranked 16 results
Returning 16 search results
INFO:     127.0.0.1:xxxxx - "POST /ragz/chat HTTP/1.1" 200 OK

Failed to get own user ID. Abort scanning.
[DEBUG] Invalid API response on page 1
[DEBUG] Finished fetching. Found 0 documents.
[DEBUG] Invalid API response on page 1

📜 Paperless-ngx Logs

🖼️ Screenshots of your settings page

No response

🖥️ Desktop Environment

Windows

💻 OS Version

Windows 11

🌐 Browser

Other

🔢 Browser Version

opera

🌐 Mobile Browser

No response

📝 Additional Information

• I have checked existing issues and this is not a duplicate
• I have tried debugging this issue on my own
• I can provide a fix and submit a PR
• I am sure that this problem is affecting everyone, not only me
• I have provided all required information above

📌 Extra Notes

No response

[oilervoss]

Same here, but I am not getting the User ID or invalid API errors.

I've set the "Process only specific pre tagged documents?" to 'ollama' tag. I have 4 documents with that tag.

See the paper-ai docker log when I hit the 'manual scan':

INFO:     Application startup complete.
INFO:     Uvicorn running on http://127.0.0.1:8000 (Press CTRL+C to quit)
[DEBUG] Found own user ID: 7
[DEBUG] Refreshing tag cache...
[DEBUG] Fetched page 1, got 6 tags. [DEBUG] Total so far: 6
[DEBUG] Tag cache refreshed. Found 6 tags.
[DEBUG] Found tag "ollama" in cache with ID 335
[DEBUG] Filtering documents for tag IDs: [ 335 ]
[DEBUG] Found own user ID: 7
[DEBUG] Fetched page 1, got 0 documents. [DEBUG] Total so far: 0
[DEBUG] Finished fetching. Found 0 documents.
[INFO] Task completed

I've checked the paperless UI and the ollama tag is indeed the 335

If I remove the "Process only specific pre tagged documents?", the paper-ai starts to process all my 160 documents.

[fmotta]

Same here:
No documents can be selected in dropdown
Choosing a tag associated with documents does not allow selection of document(s)
'Analyze with Ai' buttons are inactive

V3.0.9