I have searched for any existing and/or related issues.
I have searched for any existing and/or related discussions.
I have also searched in the CLOSED issues AND CLOSED discussions and found no related items (your issue might already be addressed on the development branch!).
I am using the latest version of Open WebUI.
Installation Method
Pip Install
Open WebUI Version
0.8.12
Ollama Version (if applicable)
No response
Operating System
Ubuntu 24.04
Browser (if applicable)
No response
Confirmation
I have read and followed all instructions in README.md.
I am using the latest version of both Open WebUI and Ollama.
I have included the browser console logs.
I have included the Docker container logs.
I have provided every relevant configuration, setting, and environment variable used in my setup.
I have clearly listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc).
I have documented step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation. My steps:
Start with the initial platform/version/OS and dependencies used,
Specify exact install/launch/configure commands,
List URLs visited, user input (incl. example values/emails/passwords if needed),
Describe all options and toggles enabled or changed,
Include any files or environmental changes,
Identify the expected and actual result at each stage,
Ensure any reasonably skilled user can follow and hit the same issue.
Expected Behavior
In channels, questions directed to the AI model should have access to the full conversation context. I guess this includes previous messages, as well as any uploaded documents or content shared in the channel.
After uploading a document or pasting its content, asking a question (and mentioning the model) should produce a response that clearly uses that context, rather than a generic answer.
Actual Behavior
In channel conversations, the model does not appear to use the provided document context. After uploading a document (PDF) and asking a question about it (while mentioning the model), the response is generic and not grounded in the document.
This also happens when the document content is manually pasted into the channel, the model still responds as if no additional context was supplied.
This does not occur in regular Chats, where uploading the same PDF and running the same test works as expected.
Steps to Reproduce
Install Open WebUI with pip install open-webui==0.8.12
Run Open WebUI with open-webui serve
Enable channels feature in Admin Panel
Create a new channel in the sidebar
Create some initial context and ask the IA model with "@"
Logs & Screenshots
Chat usage
Logs shows that Open WebUI uses query_doc and sends related context to ollama:
time=2026-04-13T13:55:31.976+02:00 level=TRACE source=server.go:1537 msg="completion request" prompt="<|start_header_id|>user<|end_header_id|>\n\n### Task:\nRespond to the user query using the provided context, incorporating inline citations in the format [id] **only when the <source> tag includes an explicit id attribute** (e.g., <source id=\"1\">).\n\n### Guidelines:\n- If you don't know the answer, clearly state that.\n- If uncertain, ask the user for clarification.\n-
Respond in the same language as the user's query.\n- If the context is unreadable or of poor quality, inform the user and provide the best possible answer.\n- If the answer isn't present in the context but you possess the knowledge, expl
ain this to the user and provide the answer using your own understanding.\n- **Only include inline citations using [id] (e.g., [1], [2]) when the <source> tag includes an id attribute.**\n- Do not cite if the <source> tag does not contain
an id attribute.\n- Do not use XML tags in your response.\n- Ensure citations are concise and directly related to the information provided.\n\n### Example of Citation:\nIf the user asks about a specific topic and the information is found in a source with a provided id attribute, the response should include the citation like in the following example:\n* \"According to the study, the proposed method increases efficiency by 20% [1].\"\n\n### Output:\nProvide a clear and dir
ect response to the user's query, including inline citations in the format [id] only when the <source> tag with id attribute is present in the context.\n\n<context>\n<source id=\"1\" name=\"JUNO_Fact_Sheet_2016.pdf\">Juno will be the first to get a detailed look at the planet's\npoles.\nJuno will take 14 days to complete each orbit, while\nJupiter spins every 10 hours. Mission planners designed</source>\n<source id=\"1\" name=\"JUNO_Fact_Sheet_2016.pdf\">SPACECRAFT DIMENSIONS\nDiameter: 66 feet (20 meters)\nHeight: 15 feet (4.5 meters)\nJovian Auroral\nDistributions\nExperiment (JADE)\nUltraviolet\nSpectrograph (UVS)\nJupiter Energetic-particle\nDetector Instrument (JEDI)\nMicrowave\nRadiometer (MWR)\nJov
ian Infrared\nAuroral Mapper\n(JIRAM)\nPlasma Waves Instrument\n(WAVES)\nLaunch August 2011\nEarth flyby October 2013\nJupiter arrival July 2016\nEnd of mission February 2018</source>\n<source id=\"1\" name=\"JUNO_Fact_Sheet_2016.pdf\">Juno Mission to Jupiter\nNational Aeronautics and Space Administration\nWith the exception of the sun, Jupiter is the most\ndominant object in the solar system. Because of its\nenormous size and the fact that it was likely the first\nof the planets to form, it has profoundly influenced the\nformation and evolution of the other bodies that orbit\nour star. NASA's Juno mission will allow us to examine\nthis gas giant planet from its innermost core to the outer\nreaches of i
ts enormous magnetic force field.\nDuring its mission, Juno will map Jupiter's gravity\nand magnetic fields to learn what the planet's interior\nstructure is like. The spacecraft also will observe the\ncomposition and circulation of the d
eep atmosphere\nand improve our understanding of the forces that\ncontrol the planet's powerful auroras. In addition\nto expanding our knowledge of the solar system's\nlargest planet, these investigations will provide clues\nabout what co
nditions in the early solar system</source>\n</context>\n\nWhat is the mission about?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n"
Channel usage
There are no query_doc calling from channels, and Ollama receives the following prompt instead:
time=2026-04-13T13:59:16.359+02:00 level=TRACE source=server.go:1537 msg="completion request" prompt="<|start_header_id|>system<|end_header_id|>\n\nYou are llama3.3:70b-instruct-q4_K_M, participating in a threaded conversation. Be concise and conversational.Here's the thread history:\n\n\nCarlos: llama3.3:70b-instruct-q4_K_M What is the mission about?\n\n\nContinue the conversation naturally as llama3.3:70b-instruct-q4_K_M, addressing the most recent message while being aware of the full context.<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nCarlos: llama3.3:70b-instruct-q4_K_M What is the mission about?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n"
Additional Information
No response
Originally created by @carpemonf on GitHub (Apr 13, 2026).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/23686
### Check Existing Issues
- [x] I have searched for any existing and/or related issues.
- [x] I have searched for any existing and/or related discussions.
- [x] I have also searched in the CLOSED issues AND CLOSED discussions and found no related items (your issue might already be addressed on the development branch!).
- [x] I am using the latest version of Open WebUI.
### Installation Method
Pip Install
### Open WebUI Version
0.8.12
### Ollama Version (if applicable)
_No response_
### Operating System
Ubuntu 24.04
### Browser (if applicable)
_No response_
### Confirmation
- [x] I have read and followed all instructions in `README.md`.
- [x] I am using the latest version of **both** Open WebUI and Ollama.
- [x] I have included the browser console logs.
- [x] I have included the Docker container logs.
- [x] I have **provided every relevant configuration, setting, and environment variable used in my setup.**
- [x] I have clearly **listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup** (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc).
- [x] I have documented **step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation**. My steps:
- Start with the initial platform/version/OS and dependencies used,
- Specify exact install/launch/configure commands,
- List URLs visited, user input (incl. example values/emails/passwords if needed),
- Describe all options and toggles enabled or changed,
- Include any files or environmental changes,
- Identify the expected and actual result at each stage,
- Ensure any reasonably skilled user can follow and hit the same issue.
### Expected Behavior
In channels, questions directed to the AI model should have access to the full conversation context. I guess this includes previous messages, as well as any uploaded documents or content shared in the channel.
After uploading a document or pasting its content, asking a question (and mentioning the model) should produce a response that clearly uses that context, rather than a generic answer.
### Actual Behavior
In channel conversations, the model does not appear to use the provided document context. After uploading a document (PDF) and asking a question about it (while mentioning the model), the response is generic and not grounded in the document.
This also happens when the document content is manually pasted into the channel, the model still responds as if no additional context was supplied.
This does not occur in regular Chats, where uploading the same PDF and running the same test works as expected.
### Steps to Reproduce
1. Install Open WebUI with `pip install open-webui==0.8.12`
2. Run Open WebUI with `open-webui serve`
3. Enable channels feature in Admin Panel
4. Create a new channel in the sidebar
5. Create some initial context and ask the IA model with "@"
### Logs & Screenshots
**Chat usage**
<img width="1915" height="990" alt="Image" src="https://github.com/user-attachments/assets/ab3051ce-abb2-454c-9a98-2f3402ce1e69" />
Logs shows that Open WebUI uses `query_doc` and sends related context to `ollama`:
```
2026-04-13 13:58:04.857 | INFO | open_webui.retrieval.utils:query_doc:157 - query_doc:result [['0f8787e6-2d12-41f3-b1f6-1847e8873529', 'a49b64bc-3ac3-4a54-93e3-8b2aac4c56e0', 'a4bacc3a-e634-4579-a624-87ac40e2daa8']] [[{'created_by': '7a3a1aba-fd1b-4604-8fd3-b4249dba266c', 'total_pages': 2, 'source': 'JUNO_Fact_Sheet_2016.pdf', 'page': 0, 'trapped': '/False', 'creationdate': '2015-09-22T10:34:10-07:00', 'producer': 'Adobe PDF Library 10.0.1', 'name': 'JUNO_Fact_Sheet_2016.pdf', 'moddate': '2016-06-13T15:25:37-07:00', 'embedding_config': "{'engine': '', 'model': 'sentence-transformers/all-MiniLM-L6-v2'}", 'start_index': 3673, 'file_id': '34e83ea9-8d49-4e48-9074-e99712940d48', 'hash': 'e52424ae2774488054947964b3729512a0569ffc62e33c3d7ff2046c14837cde', 'creator': 'Adobe InDesign CS6 (Macintosh)', 'page_label': '1'}, {'moddate': '2016-06-13T15:25:37-07:00', 'hash': 'e52424ae2774488054947964b3729512a0569ffc62e33c3d7ff2046c14837cde', 'total_pages': 2, 'page': 1, 'producer': 'Adobe PDF Library 10.0.1', 'page_label': '2', 'name': 'JUNO_Fact_Sheet_2016.pdf', 'start_index': 2671, 'trapped': '/False', 'embedding_config': "{'engine': '', 'model': 'sentence-transformers/all-MiniLM-L6-v2'}", 'creator': 'Adobe InDesign CS6 (Macintosh)', 'source': 'JUNO_Fact_Sheet_2016.pdf', 'created_by': '7a3a1aba-fd1b-4604-8fd3-b4249dba266c', 'file_id': '34e83ea9-8d49-4e48-9074-e99712940d48', 'creationdate': '2015-09-22T10:34:10-07:00'}, {'producer': 'Adobe PDF Library 10.0.1', 'hash': 'e52424ae2774488054947964b3729512a0569ffc62e33c3d7ff2046c14837cde', 'file_id': '34e83ea9-8d49-4e48-9074-e99712940d48', 'page_label': '1', 'source': 'JUNO_Fact_Sheet_2016.pdf', 'creator': 'Adobe InDesign CS6 (Macintosh)', 'page': 0, 'moddate': '2016-06-13T15:25:37-07:00', 'total_pages': 2, 'creationdate': '2015-09-22T10:34:10-07:00', 'start_index': 0, 'name': 'JUNO_Fact_Sheet_2016.pdf', 'created_by': '7a3a1aba-fd1b-4604-8fd3-b4249dba266c', 'embedding_config': "{'engine': '', 'model': 'sentence-transformers/all-MiniLM-L6-v2'}", 'trapped': '/False'}]]
2026-04-13 13:58:04.859 | INFO | open_webui.retrieval.utils:query_doc:157 - query_doc:result [['0f8787e6-2d12-41f3-b1f6-1847e8873529', 'a49b64bc-3ac3-4a54-93e3-8b2aac4c56e0', 'a4bacc3a-e634-4579-a624-87ac40e2daa8']] [[{'embedding_config': "{'engine': '', 'model': 'sentence-transformers/all-MiniLM-L6-v2'}", 'page_label': '1', 'source': 'JUNO_Fact_Sheet_2016.pdf', 'moddate': '2016-06-13T15:25:37-07:00', 'creator': 'Adobe InDesign CS6 (Macintosh)', 'hash': 'e52424ae2774488054947964b3729512a0569ffc62e33c3d7ff2046c14837cde', 'total_pages': 2, 'producer': 'Adobe PDF Library 10.0.1', 'start_index': 3673, 'name': 'JUNO_Fact_Sheet_2016.pdf', 'trapped': '/False', 'page': 0, 'file_id': '34e83ea9-8d49-4e48-9074-e99712940d48', 'created_by': '7a3a1aba-fd1b-4604-8fd3-b4249dba266c', 'creationdate': '2015-09-22T10:34:10-07:00'}, {'hash': 'e52424ae2774488054947964b3729512a0569ffc62e33c3d7ff2046c14837cde', 'moddate': '2016-06-13T15:25:37-07:00', 'created_by': '7a3a1aba-fd1b-4604-8fd3-b4249dba266c', 'file_id': '34e83ea9-8d49-4e48-9074-e99712940d48', 'creationdate': '2015-09-22T10:34:10-07:00', 'embedding_config': "{'engine': '', 'model': 'sentence-transformers/all-MiniLM-L6-v2'}", 'start_index': 2671, 'page_label': '2', 'total_pages': 2, 'producer': 'Adobe PDF Library 10.0.1', 'page': 1, 'name': 'JUNO_Fact_Sheet_2016.pdf', 'trapped': '/False', 'source': 'JUNO_Fact_Sheet_2016.pdf', 'creator': 'Adobe InDesign CS6 (Macintosh)'}, {'page_label': '1', 'moddate': '2016-06-13T15:25:37-07:00', 'trapped': '/False', 'total_pages': 2, 'start_index': 0, 'created_by': '7a3a1aba-fd1b-4604-8fd3-b4249dba266c', 'name': 'JUNO_Fact_Sheet_2016.pdf', 'creator': 'Adobe InDesign CS6 (Macintosh)', 'creationdate': '2015-09-22T10:34:10-07:00', 'source': 'JUNO_Fact_Sheet_2016.pdf', 'producer': 'Adobe PDF Library 10.0.1', 'embedding_config': "{'engine': '', 'model': 'sentence-transformers/all-MiniLM-L6-v2'}", 'file_id': '34e83ea9-8d49-4e48-9074-e99712940d48', 'hash': 'e52424ae2774488054947964b3729512a0569ffc62e33c3d7ff2046c14837cde', 'page': 0}]]
2026-04-13 13:58:04.859 | INFO | open_webui.retrieval.utils:query_doc:157 - query_doc:result [['0f8787e6-2d12-41f3-b1f6-1847e8873529', 'a49b64bc-3ac3-4a54-93e3-8b2aac4c56e0', 'a4bacc3a-e634-4579-a624-87ac40e2daa8']] [[{'creationdate': '2015-09-22T10:34:10-07:00', 'start_index': 3673, 'total_pages': 2, 'trapped': '/False', 'producer': 'Adobe PDF Library 10.0.1', 'embedding_config': "{'engine': '', 'model': 'sentence-transformers/all-MiniLM-L6-v2'}", 'created_by': '7a3a1aba-fd1b-4604-8fd3-b4249dba266c', 'creator': 'Adobe InDesign CS6 (Macintosh)', 'page_label': '1', 'hash': 'e52424ae2774488054947964b3729512a0569ffc62e33c3d7ff2046c14837cde', 'file_id': '34e83ea9-8d49-4e48-9074-e99712940d48', 'page': 0, 'moddate': '2016-06-13T15:25:37-07:00', 'source': 'JUNO_Fact_Sheet_2016.pdf', 'name': 'JUNO_Fact_Sheet_2016.pdf'}, {'page': 1, 'page_label': '2', 'creator': 'Adobe InDesign CS6 (Macintosh)', 'moddate': '2016-06-13T15:25:37-07:00', 'hash': 'e52424ae2774488054947964b3729512a0569ffc62e33c3d7ff2046c14837cde', 'embedding_config': "{'engine': '', 'model': 'sentence-transformers/all-MiniLM-L6-v2'}", 'creationdate': '2015-09-22T10:34:10-07:00', 'created_by': '7a3a1aba-fd1b-4604-8fd3-b4249dba266c', 'source': 'JUNO_Fact_Sheet_2016.pdf', 'trapped': '/False', 'start_index': 2671, 'file_id': '34e83ea9-8d49-4e48-9074-e99712940d48', 'total_pages': 2, 'producer': 'Adobe PDF Library 10.0.1', 'name': 'JUNO_Fact_Sheet_2016.pdf'}, {'producer': 'Adobe PDF Library 10.0.1', 'created_by': '7a3a1aba-fd1b-4604-8fd3-b4249dba266c', 'embedding_config': "{'engine': '', 'model': 'sentence-transformers/all-MiniLM-L6-v2'}", 'total_pages': 2, 'creator': 'Adobe InDesign CS6 (Macintosh)', 'hash': 'e52424ae2774488054947964b3729512a0569ffc62e33c3d7ff2046c14837cde', 'source': 'JUNO_Fact_Sheet_2016.pdf', 'page_label': '1', 'creationdate': '2015-09-22T10:34:10-07:00', 'page': 0, 'name': 'JUNO_Fact_Sheet_2016.pdf', 'file_id': '34e83ea9-8d49-4e48-9074-e99712940d48', 'trapped': '/False', 'start_index': 0, 'moddate': '2016-06-13T15:25:37-07:00'}]]
```
```
time=2026-04-13T13:55:31.976+02:00 level=TRACE source=server.go:1537 msg="completion request" prompt="<|start_header_id|>user<|end_header_id|>\n\n### Task:\nRespond to the user query using the provided context, incorporating inline citations in the format [id] **only when the <source> tag includes an explicit id attribute** (e.g., <source id=\"1\">).\n\n### Guidelines:\n- If you don't know the answer, clearly state that.\n- If uncertain, ask the user for clarification.\n-
Respond in the same language as the user's query.\n- If the context is unreadable or of poor quality, inform the user and provide the best possible answer.\n- If the answer isn't present in the context but you possess the knowledge, expl
ain this to the user and provide the answer using your own understanding.\n- **Only include inline citations using [id] (e.g., [1], [2]) when the <source> tag includes an id attribute.**\n- Do not cite if the <source> tag does not contain
an id attribute.\n- Do not use XML tags in your response.\n- Ensure citations are concise and directly related to the information provided.\n\n### Example of Citation:\nIf the user asks about a specific topic and the information is found in a source with a provided id attribute, the response should include the citation like in the following example:\n* \"According to the study, the proposed method increases efficiency by 20% [1].\"\n\n### Output:\nProvide a clear and dir
ect response to the user's query, including inline citations in the format [id] only when the <source> tag with id attribute is present in the context.\n\n<context>\n<source id=\"1\" name=\"JUNO_Fact_Sheet_2016.pdf\">Juno will be the first to get a detailed look at the planet's\npoles.\nJuno will take 14 days to complete each orbit, while\nJupiter spins every 10 hours. Mission planners designed</source>\n<source id=\"1\" name=\"JUNO_Fact_Sheet_2016.pdf\">SPACECRAFT DIMENSIONS\nDiameter: 66 feet (20 meters)\nHeight: 15 feet (4.5 meters)\nJovian Auroral\nDistributions\nExperiment (JADE)\nUltraviolet\nSpectrograph (UVS)\nJupiter Energetic-particle\nDetector Instrument (JEDI)\nMicrowave\nRadiometer (MWR)\nJov
ian Infrared\nAuroral Mapper\n(JIRAM)\nPlasma Waves Instrument\n(WAVES)\nLaunch August 2011\nEarth flyby October 2013\nJupiter arrival July 2016\nEnd of mission February 2018</source>\n<source id=\"1\" name=\"JUNO_Fact_Sheet_2016.pdf\">Juno Mission to Jupiter\nNational Aeronautics and Space Administration\nWith the exception of the sun, Jupiter is the most\ndominant object in the solar system. Because of its\nenormous size and the fact that it was likely the first\nof the planets to form, it has profoundly influenced the\nformation and evolution of the other bodies that orbit\nour star. NASA's Juno mission will allow us to examine\nthis gas giant planet from its innermost core to the outer\nreaches of i
ts enormous magnetic force field.\nDuring its mission, Juno will map Jupiter's gravity\nand magnetic fields to learn what the planet's interior\nstructure is like. The spacecraft also will observe the\ncomposition and circulation of the d
eep atmosphere\nand improve our understanding of the forces that\ncontrol the planet's powerful auroras. In addition\nto expanding our knowledge of the solar system's\nlargest planet, these investigations will provide clues\nabout what co
nditions in the early solar system</source>\n</context>\n\nWhat is the mission about?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n"
```
**Channel usage**
<img width="1915" height="990" alt="Image" src="https://github.com/user-attachments/assets/36e847de-ac37-4f15-86f5-f732df2c3805" />
<img width="1915" height="990" alt="Image" src="https://github.com/user-attachments/assets/8143f66d-66a8-47c2-bf2a-58a2a65a97f3" />
There are no `query_doc` calling from channels, and Ollama receives the following prompt instead:
```
time=2026-04-13T13:59:16.359+02:00 level=TRACE source=server.go:1537 msg="completion request" prompt="<|start_header_id|>system<|end_header_id|>\n\nYou are llama3.3:70b-instruct-q4_K_M, participating in a threaded conversation. Be concise and conversational.Here's the thread history:\n\n\nCarlos: llama3.3:70b-instruct-q4_K_M What is the mission about?\n\n\nContinue the conversation naturally as llama3.3:70b-instruct-q4_K_M, addressing the most recent message while being aware of the full context.<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nCarlos: llama3.3:70b-instruct-q4_K_M What is the mission about?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n"
```
### Additional Information
_No response_
GiteaMirror
added the bug label 2026-04-20 02:37:39 -05:00
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Originally created by @carpemonf on GitHub (Apr 13, 2026).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/23686
Check Existing Issues
Installation Method
Pip Install
Open WebUI Version
0.8.12
Ollama Version (if applicable)
No response
Operating System
Ubuntu 24.04
Browser (if applicable)
No response
Confirmation
README.md.Expected Behavior
In channels, questions directed to the AI model should have access to the full conversation context. I guess this includes previous messages, as well as any uploaded documents or content shared in the channel.
After uploading a document or pasting its content, asking a question (and mentioning the model) should produce a response that clearly uses that context, rather than a generic answer.
Actual Behavior
In channel conversations, the model does not appear to use the provided document context. After uploading a document (PDF) and asking a question about it (while mentioning the model), the response is generic and not grounded in the document.
This also happens when the document content is manually pasted into the channel, the model still responds as if no additional context was supplied.
This does not occur in regular Chats, where uploading the same PDF and running the same test works as expected.
Steps to Reproduce
pip install open-webui==0.8.12open-webui serveLogs & Screenshots
Chat usage
Logs shows that Open WebUI uses
query_docand sends related context toollama:Channel usage
There are no
query_doccalling from channels, and Ollama receives the following prompt instead:Additional Information
No response
@tjbck commented on GitHub (Apr 14, 2026):
#8050