[Question]: I need to call ragflow via API to query the document, and the images in the document need to be returned via links.

[zhangshenghang]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (Language Policy).
• Non-english title submitions will be closed directly ( 非英文标题的提交将会被直接关闭 ) (Language Policy).
• Please do not modify this template :) and fill in all the required fields.

Describe your problem

There are images in my PDF. Does ragflow support converting the images into links and then returning them through knowledge base querying? I need to call ragflow via API to query the document, and the images in the document need to be returned via links.

[dosubot]

Hi @zhangshenghang! I'm Dosu and I’m helping the ragflow team.

RAGFlow can extract images (such as figures) from PDFs during document parsing. These images are stored as binary blobs and assigned an img_id. When you query document chunks via the API (for example, using /v1/chunk/list), each chunk may include an img_id field if it contains an extracted image. You can then construct a link to access the image using the endpoint /v1/document/image/<img_id>, which serves the image binary directly. This allows you to retrieve images as links in your API responses if the chunk includes an image reference. The system does not extract original embedded image files (like JPEGs/PNGs) but rasterizes and crops figures from the PDF pages instead. For more details, see the implementation and discussion here and here.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}   

[zhangshenghang]

Hi @zhangshenghang! I'm Dosu and I’m helping the ragflow team.

RAGFlow can extract images (such as figures) from PDFs during document parsing. These images are stored as binary blobs and assigned an img_id. When you query document chunks via the API (for example, using /v1/chunk/list), each chunk may include an img_id field if it contains an extracted image. You can then construct a link to access the image using the endpoint /v1/document/image/<img_id>, which serves the image binary directly. This allows you to retrieve images as links in your API responses if the chunk includes an image reference. The system does not extract original embedded image files (like JPEGs/PNGs) but rasterizes and crops figures from the PDF pages instead. For more details, see the implementation and discussion here and here.

To reply, just mention @dosu.

How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other   

@dosu Which data sharding method supports this feature?

[dosubot]

RAGFlow shards image and document storage by knowledge base ID (kb_id), which maps to a storage bucket in the backend (like Minio, S3, or Azure). When images are extracted from PDFs, they're stored as binary blobs within the bucket for that knowledge base. The API endpoints (such as /v1/document/image/<img_id>) use this sharding scheme to retrieve images efficiently. This approach supports scalable, multi-tenant deployments and is consistent across all supported storage backends. You can see this logic in the storage abstraction and API implementation here and here.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}