I have searched the existing issues and discussions.
Problem Description
When uploading documents into either the Notes section or the Knowledge Library (for RAG-based search), the content is converted into Markdown. During this process, images embedded in the document are lost.
However, images are available in the payload provided by the document parser, which means they could technically be preserved in both contexts.
(At least for the Marker API and Docling as far as I know).
Currently:
In Notes, images can sometimes be pasted successfully, but this only works when copying from tools like MS Paint.
Copy-pasting directly from Word or other sources does not embed the image. This may be related to format handling (e.g., BMP vs JPEG/PNG support).
For uploaded documents, images are stripped entirely during the Markdown conversion.
Desired Solution you'd like
Preserve images from markdown conversion.
At least for notes, for knowledge/RAG this maybe needs a lot more changes to be able to add this to the database.
Improve format handling so common image formats (JPEG, PNG) are supported consistently in Notes.
Ensure uploaded documents in the Knowledge Library retain images as part of the searchable/retrievable content.
Benefits
Higher fidelity: Uploaded documents remain closer to their original form, preserving essential visual information.
Improved usability in Notes: Users can collaborate on documents that include both text and images, not just stripped-down text. Effectively making notes an AI document toolkit.
Stronger RAG capabilities: Images often carry important semantic information (e.g., diagrams, charts). Preserving them enhances the knowledge base for retrieval and reasoning.
Consistency across features: A unified approach to handling document uploads benefits both collaborative work (Notes) and knowledge management (Library).
Alternatives Considered
No response
Additional Context
No response
Originally created by @HenkieTenkie62 on GitHub (Sep 17, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/17500
### Check Existing Issues
- [x] I have searched the existing issues and discussions.
### Problem Description
When uploading documents into either the Notes section or the Knowledge Library (for RAG-based search), the content is converted into Markdown. During this process, images embedded in the document are lost.
However, images are available in the payload provided by the document parser, which means they could technically be preserved in both contexts.
(At least for the Marker API and Docling as far as I know).
**Currently:**
In Notes, images can sometimes be pasted successfully, but this only works when copying from tools like MS Paint.
Copy-pasting directly from Word or other sources does not embed the image. This may be related to format handling (e.g., BMP vs JPEG/PNG support).
For uploaded documents, images are stripped entirely during the Markdown conversion.
### Desired Solution you'd like
- Preserve images from markdown conversion.
At least for notes, for knowledge/RAG this maybe needs a lot more changes to be able to add this to the database.
- Improve format handling so common image formats (JPEG, PNG) are supported consistently in Notes.
- Ensure uploaded documents in the Knowledge Library retain images as part of the searchable/retrievable content.
Benefits
Higher fidelity: Uploaded documents remain closer to their original form, preserving essential visual information.
Improved usability in Notes: Users can collaborate on documents that include both text and images, not just stripped-down text. Effectively making notes an AI document toolkit.
Stronger RAG capabilities: Images often carry important semantic information (e.g., diagrams, charts). Preserving them enhances the knowledge base for retrieval and reasoning.
Consistency across features: A unified approach to handling document uploads benefits both collaborative work (Notes) and knowledge management (Library).
### Alternatives Considered
_No response_
### Additional Context
_No response_
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 @HenkieTenkie62 on GitHub (Sep 17, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/17500
Check Existing Issues
Problem Description
When uploading documents into either the Notes section or the Knowledge Library (for RAG-based search), the content is converted into Markdown. During this process, images embedded in the document are lost.
However, images are available in the payload provided by the document parser, which means they could technically be preserved in both contexts.
(At least for the Marker API and Docling as far as I know).
Currently:
In Notes, images can sometimes be pasted successfully, but this only works when copying from tools like MS Paint.
Copy-pasting directly from Word or other sources does not embed the image. This may be related to format handling (e.g., BMP vs JPEG/PNG support).
For uploaded documents, images are stripped entirely during the Markdown conversion.
Desired Solution you'd like
At least for notes, for knowledge/RAG this maybe needs a lot more changes to be able to add this to the database.
Benefits
Higher fidelity: Uploaded documents remain closer to their original form, preserving essential visual information.
Improved usability in Notes: Users can collaborate on documents that include both text and images, not just stripped-down text. Effectively making notes an AI document toolkit.
Stronger RAG capabilities: Images often carry important semantic information (e.g., diagrams, charts). Preserving them enhances the knowledge base for retrieval and reasoning.
Consistency across features: A unified approach to handling document uploads benefits both collaborative work (Notes) and knowledge management (Library).
Alternatives Considered
No response
Additional Context
No response