mirror of
https://github.com/open-webui/open-webui.git
synced 2026-07-15 21:19:39 -05:00
[GH-ISSUE #14807] issue: Error calling Docling: Error calling Docling API: Not Found - Task result not found. Please wait for a completion status. #17369
No Assignees
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: github-starred/open-webui#17369
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @gongshaojie12 on GitHub (Jun 9, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/14807
Check Existing Issues
Installation Method
Pip Install
Open WebUI Version
0.6.13
Ollama Version (if applicable)
No response
Operating System
centos7.9
Browser (if applicable)
No response
Confirmation
README.md.Expected Behavior
Files can be uploaded normally.
Actual Behavior
Whenever I try to upload files larger than 1MB in the dialog box, I always encounter the following issue:
2025-06-09 17:04:47.137 | ERROR | open_webui.routers.retrieval:process_file:1413 - Error calling Docling: Error calling Docling API: Not Found - Task result not found. Please wait for a completion status. - {}Docling logs as follows:
ERROR:docling_serve.engines.async_local.worker:Worker 0 failed to process job 8e75bace-451e-4898-ac3f-9f11f2c51db1: 500:But I can upload successfully through the Docling UI.
How can I solve this problem? Thanks!
Steps to Reproduce
None
Logs & Screenshots
None
Additional Information
No response
@Expro commented on GitHub (Jun 10, 2025):
Seems like open-webui is incompatible with docling-serve v12 and v13. Last working version is v11.
@maxdcy commented on GitHub (Jun 11, 2025):
use this https://github.com/docling-project/docling-serve/pkgs/container/docling-serve-cu124/422590280?tag=v0.11.0
and got it working for now until the devs sort it out :)
@geiseri commented on GitHub (Jun 12, 2025):
I could only get it to work if I disable ocr. Even with v0.11.0, it complains about the language options.
@vaclcer commented on GitHub (Jun 17, 2025):
Is this still a thing? Can you re-test with OWU v0.6.15 and docling-serve v0.13.0?
@kunom commented on GitHub (Jun 17, 2025):
Here, Docling seems to crash with:
ERROR:docling_serve.engines.async_local.worker:Worker 0 failed to process job ca8fb8a2-b3fc-4e77-a733-5f7589c4e3d2: 500: Repo id must be in the form 'repo_name' or 'namespace/repo_name': '/opt/app-root/src/.cache/docling/models/HuggingFaceTB--SmolVLM-256M-Instruct'. Userepo_typeargument if needed.Direct conversion of the file through the Docling UI (standard settings) works.
@ginochen-avalue commented on GitHub (Jun 18, 2025):
I’m having the same problem.
Docling Log
ERROR:docling_serve.engines.async_local.worker:Worker 0 failed to process job 87c59b87-f03b-4f3e-aafb-f39dd00bf00e: 500: Repo id must be in the form 'repo_name' or 'namespace/repo_name': '/opt/app-root/src/.cache/docling/models/HuggingFaceTB--SmolVLM-256M-Instruct'. Use
repo_typeargument if needed.INFO: 172.18.0.1:39070 - "POST /v1alpha/convert/file HTTP/1.1" 404 Not Found
OpenWebUI Log
2025-06-18 09:22:47.370 | INFO | open_webui.routers.files:upload_file:94 - file.content_type: application/pdf - {}
2025-06-18 09:22:52.392 | ERROR | open_webui.routers.retrieval:process_file:1457 - Error calling Docling: Error calling Docling API: Not Found - Task result not found. Please wait for a completion status. - {}
Traceback (most recent call last):
File "/usr/local/lib/python3.11/threading.py", line 1002, in _bootstrap
self._bootstrap_inner()
│ └ <function Thread._bootstrap_inner at 0x72f7b8cc89a0>
└ <WorkerThread(AnyIO worker thread, started 126404972766912)>
File "/usr/local/lib/python3.11/threading.py", line 1045, in _bootstrap_inner
self.run()
│ └ <function WorkerThread.run at 0x72f6f82e8d60>
└ <WorkerThread(AnyIO worker thread, started 126404972766912)>
File "/usr/local/lib/python3.11/site-packages/anyio/_backends/_asyncio.py", line 967, in run
result = context.run(func, *args)
│ │ │ └ ()
│ │ └ functools.partial(<function upload_file at 0x72f7597cac00>, user=UserModel(id='0897a0bc-8b45-4aaf-a933-30003e84dc77', name='G...
│ └ <method 'run' of '_contextvars.Context' objects>
└ <_contextvars.Context object at 0x72f6f07627c0>
File "/app/backend/open_webui/routers/files.py", line 181, in upload_file
process_file(request, ProcessFileForm(file_id=id), user=user)
│ │ │ │ └ UserModel(id='0897a0bc-8b45-4aaf-a933-30003e84dc77', name='', email='', role='admin', profile_ima...
│ │ │ └ '0cdccebc-39e6-4d39-af31-cf9b49452d81'
│ │ └ <class 'open_webui.routers.retrieval.ProcessFileForm'>
│ └ <starlette.requests.Request object at 0x72f6c244e590>
└ <function process_file at 0x72f7570ed620>
File "/app/backend/open_webui/retrieval/loaders/main.py", line 218, in load
docs = loader.load()
│ └ <function DoclingLoader.load at 0x72f7582d7ce0>
└ <open_webui.retrieval.loaders.main.DoclingLoader object at 0x72f6d2649550>
File "/app/backend/open_webui/retrieval/loaders/main.py", line 206, in load
raise Exception(f"Error calling Docling: {error_msg}")
Exception: Error calling Docling: Error calling Docling API: Not Found - Task result not found. Please wait for a completion status.
2025-06-18 09:22:52.393 | ERROR | open_webui.routers.files:upload_file:190 - 400: Error calling Docling: Error calling Docling API: Not Found - Task result not found. Please wait for a completion status. - {}
Traceback (most recent call last):
File "/app/backend/open_webui/routers/retrieval.py", line 1376, in process_file
docs = loader.load(
│ └ <function Loader.load at 0x72f7582d7e20>
└ <open_webui.retrieval.loaders.main.Loader object at 0x72f6f82510d0>
File "/app/backend/open_webui/retrieval/loaders/main.py", line 218, in load
docs = loader.load()
│ └ <function DoclingLoader.load at 0x72f7582d7ce0>
└ <open_webui.retrieval.loaders.main.DoclingLoader object at 0x72f6d2649550>
File "/app/backend/open_webui/retrieval/loaders/main.py", line 206, in load
raise Exception(f"Error calling Docling: {error_msg}")
Exception: Error calling Docling: Error calling Docling API: Not Found - Task result not found. Please wait for a completion status.
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/usr/local/lib/python3.11/threading.py", line 1002, in _bootstrap
self._bootstrap_inner()
│ └ <function Thread._bootstrap_inner at 0x72f7b8cc89a0>
└ <WorkerThread(AnyIO worker thread, started 126404972766912)>
File "/usr/local/lib/python3.11/threading.py", line 1045, in _bootstrap_inner
self.run()
│ └ <function WorkerThread.run at 0x72f6f82e8d60>
└ <WorkerThread(AnyIO worker thread, started 126404972766912)>
File "/usr/local/lib/python3.11/site-packages/anyio/_backends/_asyncio.py", line 967, in run
result = context.run(func, *args)
│ │ │ └ ()
│ │ └ functools.partial(<function upload_file at 0x72f7597cac00>, user=UserModel(id='0897a0bc-8b45-4aaf-a933-30003e84dc77', name='G...
│ └ <method 'run' of '_contextvars.Context' objects>
└ <_contextvars.Context object at 0x72f6f07627c0>
File "/app/backend/open_webui/routers/retrieval.py", line 1464, in process_file
raise HTTPException(
└ <class 'fastapi.exceptions.HTTPException'>
fastapi.exceptions.HTTPException: 400: Error calling Docling: Error calling Docling API: Not Found - Task result not found. Please wait for a completion status.
2025-06-18 09:22:52.394 | ERROR | open_webui.routers.files:upload_file:191 - Error processing file: 0cdccebc-39e6-4d39-af31-cf9b49452d81 - {}
2025-06-18 09:22:52.397 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 192.168.99.13:10027 - "POST /api/v1/files/ HTTP/1.1" 200 - {}
2025-06-18 09:22:52.407 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 192.168.99.13:10027 - "POST /api/v1/knowledge/05c7a397-5f69-4269-bafc-22c798010e4c/file/add HTTP/1.1" 400 -
@Schwenn2002 commented on GitHub (Jun 26, 2025):
Please implement support for the new doccling-serve version! The current version is 0.16.0!
@vaclcer commented on GitHub (Jun 26, 2025):
The current implementation is docling-serve v0.16.0 compatible. Before using exclamation marks, I would suggest consulting documentation:
https://docs.openwebui.com/features/document-extraction/docling/#optional-step-3-configure-doclings-picture-description-features
@bursica commented on GitHub (Jun 26, 2025):
📘 Workaround suggestion at Integrating Docling with a Local Vision LLM using Flask Proxy
This guide explains how to configure Docling to describe images using a local vision LLM model (e.g.,
llama3-vision) via a Flask proxy adapter. This enables full offline PDF/image OCR + captioning, useful for AI document processing pipelines.✅ 1. Prerequisites
eng,slv)llama3:vision)flask,requests,Pillow, etc. in your Conda environment📁 2. Project Structure
🧠 3. Flask Proxy Server Setup
proxy_adapter/app.pyproxy_adapter/requirements.txt🧪 4. Start Flask Proxy Locally
You should now have an LLM proxy listening on
http://localhost:8001.🔧 5.
docling/config.yamlConfiguration🛠️ 6. Making Flask Persistent (Optional)
Create
start_proxy_adapter.shCreate
flask.servicefor systemdInstall the service:
Check with:
✅ 7. Running Docling with the Proxy
Use the official
doclingcontainer with yourconfig.yamlmounted. Once everything is wired:Ensure
result.jsoncontains:md_content→ Full OCR outputimage_descriptions[]→ Vision model captions🧪 8. Test & Debug
If image captions don’t appear:
proxy_adapter/app.pylogcurl localhost:11434/api/generateslv.traineddatais in/usr/share/tesseract/tessdata/📦 9. Sample Result Output
🔌 10. OI Integration with Local Docling + Flask Proxy
Happy Coding 😎
@Schwenn2002 commented on GitHub (Jun 26, 2025):
I'm sorry for the emotional reaction, but even with image description disabled, it still runs into an error (as described above). For example, dociling-serve works with Word files, but not with scanned PDFs that already contain OCR text. Docling-serve version 0.11 runs without OCR on the same PDF documents. So something must have changed in the API, and open-webui is incompatible.
@bursica commented on GitHub (Jun 26, 2025):
I have gone through same emotional rollercoaster.. this workaround with flask is ok for small pdfs. still have issues with huge documents with lots of images. however loading docx etc, id doesn't run into errors, but models are unable to work with images in them, at least what I tested.
@Schwenn2002 commented on GitHub (Jun 26, 2025):
In version 0.11.0, the connection works flawlessly without OCR with the same documents. The API appears to have changed.
@vaclcer commented on GitHub (Jun 26, 2025):
Guys, I prepared that PR and I have been using this feature with OWU v0.6.15 and Docling (all the versions) since then.
This is a configuration issue.
Look at the documentation, make sure the configuration JSON object is valid. There is nothing more to say. It just works..
@Schwenn2002 commented on GitHub (Jun 26, 2025):
Then I'm actually missing the logic for how to configure it. My config works with version 0.11.0, but not with newer versions. I configured everything using docker-compose, both doclin-serve and open-webui. Could you perhaps provide a working example of the config? What else do I need to pay attention to?
@Schwenn2002 commented on GitHub (Jun 26, 2025):
Would it be possible to provide a Docker configuration for dociling-serve that actually works? Despite the instructions, I can only get a working solution with dociling-serve 0.11.0 (without OCR). What am I doing wrong?
docker-compose.yaml
@vaclcer commented on GitHub (Jun 27, 2025):
Sure. This works for me with the OWU configuration pictured at the docs: https://docs.openwebui.com/features/document-extraction/docling/#optional-step-3-configure-doclings-picture-description-features
@funnycups commented on GitHub (Jun 27, 2025):
I believe
DOCLING_SERVE_ENABLE_REMOTE_SERVICESor some other environment variables provided in docker compose by @vaclcer is necessary. Problems disappear right after I add these environment variables.I double check Docling configuration docs and found the description of
DOCLING_SERVE_ENABLE_REMOTE_SERVICES:It was actually stated in the Docling usage docs that:
I was so careless that I neglected this note. I would recommend anyone who face problems on Docling with Picture description options try adding
DOCLING_SERVE_ENABLE_REMOTE_SERVICESenvironment variable!@Expro commented on GitHub (Jun 27, 2025):
I had this environment variable set from start. Didn't help, anything above v11 still didn't work.
@vaclcer commented on GitHub (Jun 27, 2025):
Could you post a screenshot of your settings in OWU -> Documents?
@Schwenn2002 commented on GitHub (Jun 27, 2025):
I suspect that the single quotes in my environment variables were the wrong syntax, but the configuration is working for me now.
@mys721tx commented on GitHub (Jul 4, 2025):
I've made a PR (open-webui/docs#605) for some JSON examples of the docling OCR setting.
@geoHeil commented on GitHub (Jul 20, 2025):
https://github.com/open-webui/open-webui/issues/14807#issuecomment-3008386545 is there any planned upgrade to the API changes of 1.0.0?
@funnycups commented on GitHub (Jul 24, 2025):
Indeed, looks like starting from 1.0.0, docling use v1 endpoint instead of v1alpha, which leads to 404 error when using it in open webui.
There is already a draft pr of this change, https://github.com/open-webui/open-webui/pull/15785.
@Kunal-Shah107 commented on GitHub (Jul 27, 2025):
I am getting performance issue even if running on big GPU machine, I am sure something is missing.
Here is my setup
VM - Standard NV12ads A10 v5 (12 vcpus, 110 GiB memory) (Azure)
Docker Container - docker run -d --gpus all --name docling -p 5001:5001 -e DOCLING_SERVE_ENABLE_UI=true -e DOCLING_SERVE_MAX_SYNC_WAIT=600 --restart unless-stopped quay.io/docling-project/docling-serve-cu124
API Call - curl --location 'http://:5001/v1alpha/convert/file'
--form 'files=@"c:\workspace\EN.pdf"'
--form 'ocr_engine="easyocr"'
--form 'ocr_language="en,ar"'
--form 'to_formats="md"'
--form 'pdf_backend="pypdfium2"'
Even after having GPU why it is taking a longer, Can some one assist here ! Thanks in advance.
@geoHeil commented on GitHub (Jul 27, 2025):
What is long? Can you wuantify this?
Is this longer than a local (python, by hand) docling conversion on the same machine?
@Kunal-Shah107 commented on GitHub (Jul 27, 2025):
@geoHeil
I tried local as well on Standard NV12ads A10 v5 (12 vcpus, 110 GiB memory) (Azure)
I got OOM error. PDF is 7 MB
test_docling_local.py
from docling.document_converter import DocumentConverter
input_path = "/opt/test.pdf" # Change to your file location
converter = DocumentConverter()
results = converter.convert_all([input_path])
for result in results:
print("File:", result.input.file)
print("Status:", result.status)
# Check for markdown content (many results have this attr, some may not)
if hasattr(result, "md_content"):
print("Markdown output (truncated):")
print(result.md_content[:1000]) # Print first 1000 chars
else:
# If not, just dump all attributes
print(dir(result))
This for one user, imagine having multiple user uploading PDF at the same time.
How to tackle it and improve performance ?
Correct me If I am using docling the wrong way ! (newbee)
@vaclcer commented on GitHub (Jul 27, 2025):
Yes, Docling is slow and memory hungry... but you should really complain at Docling's repo, not here.
@Kunal-Shah107 commented on GitHub (Jul 27, 2025):
@vaclcer - I was looking if someone is having similar issue while integrating with OpenWebUI. Its not complaining, I will keep you posted once I hear from docling. Have a great rest of your productive day ! Cheers.
@TRKNOVA commented on GitHub (Jul 29, 2025):
hi, this is why i didn't work... give me my hair back!
@rober-tm commented on GitHub (Aug 1, 2025):
Este bug parece que se corrigira en la proxima actualizacion, hay un commit sobre ello en la rama dev, con esto
Commit
4bc054abackend/open_webui/retrieval/loaders/main.py
`
line 184
`
@arty-hlr commented on GitHub (Aug 3, 2025):
I also get:
even with
and I also can't figure what the issue is. Latest docling and open-webui dev docker containers.
@emorgan33 commented on GitHub (Aug 6, 2025):
From Perplexity AI:
The "172.17.0.1:54186 - 'POST /v1alpha/convert/file HTTP/1.1' 404 Not Found" error is caused by the Docling API no longer supporting the /v1alpha/convert/file endpoint.
Starting from Docling version 1.0.1, the endpoint path has changed to /v1/convert/file instead of /v1alpha/convert/file. So any calls using the old /v1alpha/convert/file path result in a 404 Not Found error because the route does not exist anymore.
@funnycups commented on GitHub (Aug 7, 2025):
PR about this endpoint change has already been merged. I believe next release would just fix this problem.
For now you can try temporarily downgrading Docling or building OpenWebUI yourself.
@adhusch commented on GitHub (Aug 7, 2025):
That's it.
I would be great when open webui would support specifying the full endpoint URL in the user interface, having the user specifying only the (host) prefix part of the URL is quite convenient but adds this dependency on being forced to closely follow up on external API changes. Convenience could be retained by pre-populating the full URL string including the "typical" API URL postfix, as it is currently hard coded.
@dojoca commented on GitHub (Aug 31, 2025):
Been tearing my hair out with trying to get RAG working in OpenWebUI.
@98h398hrpohpoai commented on GitHub (Sep 1, 2025):
The same error is being produced with the latest build of Open-Webui 0.6.26 w/ the corrected api patch and Docling v1.31. Downgrading docling no longer works due to the older api.
Edit: This error only occurs for me when using "describe picture" option. Resolved with removing that.
@dmanresa-saes commented on GitHub (Sep 3, 2025):
Same here
@Krashnicov commented on GitHub (Sep 9, 2025):
Came across this issue today, is there eta on fixing the describe picture issue? Many of my docs include images and losing that context is sub-optimal.
@vaclcer commented on GitHub (Sep 9, 2025):
Unless something changed in recent releases (I doubt it), "describe picture" feature works as expected - in my case OWU v0.6.22 and Docling v1.4.1
I can only suggest to look into OWU Docs closely and paying attention to the syntax.. It just works.
@Krashnicov commented on GitHub (Sep 9, 2025):
Using docling-serve:latest (pulled today).
@98h398hrpohpoai commented on GitHub (Sep 12, 2025):
Can you share your config? I've tried several and I get the error from the above post regarding task result not found (either default or local). Otherwise working with Docling v1.5 on v0.6.28
@geoHeil commented on GitHub (Sep 18, 2025):
I observe something else for OWUI 0.30 (current latest)
for me silently failing now for
works fine for
Is there a more recent confirmed working version of docling?
@artemis15 commented on GitHub (Oct 2, 2025):
I'm still getting the issue, even downgrading docling won't help.
@Ithrial commented on GitHub (Oct 3, 2025):
I just updated to docling 2.55.0 with OWUI version 6.32 and getting API errors because the new version of Docling is expecting a repo_id field and its not being passed.
@geoHeil commented on GitHub (Oct 8, 2025):
https://github.com/open-webui/open-webui/pull/17363 does not look like it is fixing the docling version
@physicalit commented on GitHub (Oct 21, 2025):
This issue is still present if Describe Picture is activated
@normen commented on GitHub (Oct 26, 2025):
I see the same, trying to use an API endpoint as described here: https://docs.openwebui.com/features/rag/document-extraction/docling#verifying-docling-in-docker
@enving commented on GitHub (Nov 11, 2025):
also the same here..even worse. even with picture describe off i will get the error:
"Error calling Docling: Error calling Docling API: Not Found - Task result not found. Please wait for a completion status."
settings and installation exactly like officially described (i hope... a bit added to call my picture ...but lie said its still deactivated anyway):
docling-serve:
image: quay.io/docling-project/docling-serve
container_name: docling-serve
restart: always
ports:
- "5001:5001"
networks:
- xx-network
volumes:
- docling-data:/app/data
environment:
- DOCLING_SERVE_ENABLE_UI=true
# Korrigierte JSON-Syntax als YAML-String:
- DOCLING_VLM_PIPELINE_MODEL_LOCAL=${DOCLING_VLM_PIPELINE_MODEL_LOCAL}
[https://docs.openwebui.com/features/rag/document-extraction/docling#verifying-docling-in-docker]
@artemis15 commented on GitHub (Nov 16, 2025):
it is solved yet?
@qhaas commented on GitHub (Nov 21, 2025):
No
@grivatyBox commented on GitHub (Nov 25, 2025):
I checked this document and followed the steps, and it solved my problem.https://github.com/docling-project/docling-serve/blob/main/docs/models.md
@Classic298 commented on GitHub (Dec 30, 2025):
The core issue (API endpoint change from
/v1alpha/convert/fileto/v1/convert/file) has been fixed in the Open WebUI codebase.For users still experiencing issues:
See docs: https://docs.openwebui.com/features/rag/document-extraction/docling
Closing as the Open WebUI fix is complete.