[Bug?] Errors when using paperless with https+own domain, no errors when using http+ip

[FHeilmann]

Setup

• paperless-ngx running in my home network, reachable under its own IP at port 80
• paperless-gpt (commit 32cc3d2 running on my personal computer for testing)
• traefik reverse proxy running in my home network
• own domain with letsencrypt certificate, any requests to any subdomain are pointed at my traefik instance

Issue description

I'm currently trying to set up paperless-gpt in my homelab and have noticed the following issues:

When I run paperless-gpt while pointing it at the domain of my paperless installation like so:

docker run -d \
  -e PAPERLESS_BASE_URL='https://paperless.mydomain' \
  -e PAPERLESS_API_TOKEN='mytoken' \
  -e LLM_PROVIDER='ollama' \
  -e LLM_MODEL='llama2' \
  -e AUTO_TAG='PaperlessGPTTagAuto' \
  -e MANUAL_TAG='PaperlessGPTTagManual' \
  -e MANUAL_OCR_TAG='PaperlessGptOCR' \
  -e LLM_LANGUAGE='English' \
  -e LOG_LEVEL='debug' \
  -p 8080:8080 \
  paperless-gpt

paperless will not start up correctly and spam the log with the following error messages:

2025-01-30 11:27:59 time="2025-01-30T10:27:59Z" level=error msg="Error fetching documents: invalid character '<' looking for beginning of value"

2025-01-30 11:27:57 time="2025-01-30T10:27:57Z" level=error msg="Error in processAutoTagDocuments: error in processAutoTagDocuments: error fetching documents with autoTag: invalid character '<' looking for beginning of value"

If I instead run paperless-gpt with the following command, pointing it at the IP of my paperless-ngx instance and using http instead of the subdomain with https:

docker run -d \
  -e PAPERLESS_BASE_URL='http://<paperless_ip>' \
  -e PAPERLESS_API_TOKEN='mytoken' \
  -e LLM_PROVIDER='ollama' \
  -e LLM_MODEL='llama2' \
  -e AUTO_TAG='PaperlessGPTTagAuto' \
  -e MANUAL_TAG='PaperlessGPTTagManual' \
  -e MANUAL_OCR_TAG='PaperlessGptOCR' \
  -e LLM_LANGUAGE='English' \
  -e LOG_LEVEL='debug' \
  -p 8080:8080 \
  paperless-gpt

paperless-gpt will start up correctly without any errors. I assume the < error stems from something trying to parse <nil> and failing. Has anyone successfully managed to run paperless-gpt against a paperless-ngx server behind a custom domain with SSL certificate?

Thanks for your time and consideration!

Addendum: Full startup logs of paperless GPT for the two cases:

With https + domain

2025-01-30 11:37:42 Using PaperlessGPTTagManual as manual tag
2025-01-30 11:37:42 Using PaperlessGPTTagAuto as auto tag
2025-01-30 11:37:42 
2025-01-30 11:37:42     ╔═══════════════════════════════════════╗
2025-01-30 11:37:42     ║             Paperless GPT             ║
2025-01-30 11:37:42     ╚═══════════════════════════════════════╝
2025-01-30 11:37:42 
2025-01-30 11:37:42 Version: docker-dev
2025-01-30 11:37:42 Commit: unknown
2025-01-30 11:37:42 Build Date: unknown
2025-01-30 11:37:42 Platform: linux/amd64
2025-01-30 11:37:42 Go Version: go1.23.4
2025-01-30 11:37:42 Started: Thu, 30 Jan 2025 10:37:42 UTC
2025-01-30 11:37:42 
2025-01-30 11:37:42 time="2025-01-30T10:37:42Z" level=info msg="Worker 0 started"
2025-01-30 11:37:42 [GIN] 2025/01/30 - 10:37:42 | 500 |   91.251526ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:37:42 [GIN] 2025/01/30 - 10:37:42 | 500 |  123.753967ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:37:42 [GIN] 2025/01/30 - 10:37:42 | 500 |   43.059463ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:37:42 [GIN] 2025/01/30 - 10:37:42 | 500 |   42.112305ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:37:43 [GIN] 2025/01/30 - 10:37:43 | 500 |   46.089673ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:37:42 time="2025-01-30T10:37:42Z" level=error msg="Could not read prompts/title_prompt.tmpl, using default template: open prompts/title_prompt.tmpl: no such file or directory"
2025-01-30 11:37:42 time="2025-01-30T10:37:42Z" level=error msg="Could not read prompts/tag_prompt.tmpl, using default template: open prompts/tag_prompt.tmpl: no such file or directory"
2025-01-30 11:37:42 time="2025-01-30T10:37:42Z" level=error msg="Could not read prompts/correspondent_prompt.tmpl, using default template: open prompts/correspondent_prompt.tmpl: no such file or directory"
2025-01-30 11:37:42 time="2025-01-30T10:37:42Z" level=error msg="Could not read prompts/ocr_prompt.tmpl, using default template: open prompts/ocr_prompt.tmpl: no such file or directory"
2025-01-30 11:37:42 time="2025-01-30T10:37:42Z" level=info msg="Vision LLM not enabled"
2025-01-30 11:37:42 time="2025-01-30T10:37:42Z" level=info msg="Server started on interface :8080"
2025-01-30 11:37:42 time="2025-01-30T10:37:42Z" level=error msg="Error in processAutoTagDocuments: error in processAutoTagDocuments: error fetching documents with autoTag: invalid character '<' looking for beginning of value"
2025-01-30 11:37:42 time="2025-01-30T10:37:42Z" level=error msg="Error fetching documents: invalid character '<' looking for beginning of value"
2025-01-30 11:37:42 time="2025-01-30T10:37:42Z" level=error msg="Error fetching documents: invalid character '<' looking for beginning of value"
2025-01-30 11:37:42 time="2025-01-30T10:37:42Z" level=error msg="Error fetching documents: invalid character '<' looking for beginning of value"
2025-01-30 11:37:42 time="2025-01-30T10:37:42Z" level=error msg="Error fetching documents: invalid character '<' looking for beginning of value"
2025-01-30 11:37:43 time="2025-01-30T10:37:43Z" level=error msg="Error fetching documents: invalid character '<' looking for beginning of value"
2025-01-30 11:37:44 time="2025-01-30T10:37:44Z" level=error msg="Error fetching documents: invalid character '<' looking for beginning of value"
2025-01-30 11:37:44 [GIN] 2025/01/30 - 10:37:44 | 500 |   42.826563ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:37:45 time="2025-01-30T10:37:45Z" level=error msg="Error fetching documents: invalid character '<' looking for beginning of value"
2025-01-30 11:37:45 [GIN] 2025/01/30 - 10:37:45 | 500 |   49.084046ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:37:46 time="2025-01-30T10:37:46Z" level=error msg="Error fetching documents: invalid character '<' looking for beginning of value"
2025-01-30 11:37:46 [GIN] 2025/01/30 - 10:37:46 | 500 |   43.083048ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:37:47 [GIN] 2025/01/30 - 10:37:47 | 500 |     44.9398ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:37:47 time="2025-01-30T10:37:47Z" level=error msg="Error fetching documents: invalid character '<' looking for beginning of value"

With http + IP

2025-01-30 11:38:34 Using PaperlessGPTTagManual as manual tag
2025-01-30 11:38:34 Using PaperlessGPTTagAuto as auto tag
2025-01-30 11:38:34 
2025-01-30 11:38:34     ╔═══════════════════════════════════════╗
2025-01-30 11:38:34     ║             Paperless GPT             ║
2025-01-30 11:38:34     ╚═══════════════════════════════════════╝
2025-01-30 11:38:34 
2025-01-30 11:38:34 Version: docker-dev
2025-01-30 11:38:34 Commit: unknown
2025-01-30 11:38:34 Build Date: unknown
2025-01-30 11:38:34 Platform: linux/amd64
2025-01-30 11:38:34 Go Version: go1.23.4
2025-01-30 11:38:34 Started: Thu, 30 Jan 2025 10:38:34 UTC
2025-01-30 11:38:34 
2025-01-30 11:38:34 time="2025-01-30T10:38:34Z" level=info msg="Worker 0 started"
2025-01-30 11:38:34 [GIN] 2025/01/30 - 10:38:34 | 200 |  186.633423ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:38:34 [GIN] 2025/01/30 - 10:38:34 | 200 |  217.794115ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:38:34 [GIN] 2025/01/30 - 10:38:34 | 200 |  197.850198ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:38:34 [GIN] 2025/01/30 - 10:38:34 | 200 |  151.986734ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:38:34 [GIN] 2025/01/30 - 10:38:34 | 200 |  124.977313ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:38:35 [GIN] 2025/01/30 - 10:38:35 | 200 |  118.355398ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:38:36 [GIN] 2025/01/30 - 10:38:36 | 200 |  117.461745ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:38:34 time="2025-01-30T10:38:34Z" level=error msg="Could not read prompts/title_prompt.tmpl, using default template: open prompts/title_prompt.tmpl: no such file or directory"
2025-01-30 11:38:34 time="2025-01-30T10:38:34Z" level=error msg="Could not read prompts/tag_prompt.tmpl, using default template: open prompts/tag_prompt.tmpl: no such file or directory"
2025-01-30 11:38:34 time="2025-01-30T10:38:34Z" level=error msg="Could not read prompts/correspondent_prompt.tmpl, using default template: open prompts/correspondent_prompt.tmpl: no such file or directory"
2025-01-30 11:38:34 time="2025-01-30T10:38:34Z" level=error msg="Could not read prompts/ocr_prompt.tmpl, using default template: open prompts/ocr_prompt.tmpl: no such file or directory"
2025-01-30 11:38:37 [GIN] 2025/01/30 - 10:38:37 | 200 |   116.27501ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:38:38 [GIN] 2025/01/30 - 10:38:38 | 200 |   118.82812ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:38:39 [GIN] 2025/01/30 - 10:38:39 | 200 |  110.982298ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:38:40 [GIN] 2025/01/30 - 10:38:40 | 200 |  116.792172ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:38:42 [GIN] 2025/01/30 - 10:38:42 | 200 |   116.19173ms |      172.17.0.1 | GET      "/api/documents"
2025-01-30 11:38:43 [GIN] 2025/01/30 - 10:38:43 | 200 |  115.853105ms |      172.17.0.1 | GET      "/api/documents"

[mkrinke]

this fixed it for me: #230