[GH-ISSUE #9629] Attachment support #52796

Closed
opened 2026-04-29 00:54:32 -05:00 by GiteaMirror · 3 comments
Owner

Originally created by @lukehutch on GitHub (Mar 10, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/9629

LLMs are increasingly supporting attachments (both Anthropic and OpenAI now support these).

I don't know if any of the current ollama models can even work with attachments, short of including them in the context window (if they are plain text, at least). However, I'm sure in the future many models will support working with attachments (including more complex formats like images or PDFs).

I don't see a way to add attachments through the ollama commandline interface. If this functionality does not currently exist, please consider this a feature request.

Originally created by @lukehutch on GitHub (Mar 10, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/9629 LLMs are increasingly supporting attachments (both Anthropic and OpenAI now support these). I don't know if any of the current ollama models can even work with attachments, short of including them in the context window (if they are plain text, at least). However, I'm sure in the future many models will support working with attachments (including more complex formats like images or PDFs). I don't see a way to add attachments through the ollama commandline interface. If this functionality does not currently exist, please consider this a feature request.
GiteaMirror added the feature request label 2026-04-29 00:54:32 -05:00
Author
Owner

@rick-github commented on GitHub (Mar 10, 2025):

Attachment processing is usually a function of the client. See integrations for clients that handle documents.

<!-- gh-comment-id:2711511842 --> @rick-github commented on GitHub (Mar 10, 2025): Attachment processing is usually a function of the client. See [integrations](https://github.com/ollama/ollama?tab=readme-ov-file#community-integrations) for clients that handle documents.
Author
Owner

@hybra commented on GitHub (Mar 11, 2025):

I am able to attach images to Llava with a prompt like

Describe this image /Macdrive/local_image_path_on_mac/image.png

(just write "Describe this image " then drag the image from the Finder to the terminal, be aware of the space after the word "image ")

Ollama will say it added the image to the prompt and Llava starts processing.

<!-- gh-comment-id:2715144591 --> @hybra commented on GitHub (Mar 11, 2025): I am able to attach images to Llava with a prompt like Describe this image /Macdrive/local_image_path_on_mac/image.png (just write "Describe this image " then drag the image from the Finder to the terminal, be aware of the space after the word "image ") Ollama will say it added the image to the prompt and Llava starts processing.
Author
Owner

@rick-github commented on GitHub (Mar 11, 2025):

Yes, image models will accept images.

<!-- gh-comment-id:2715150137 --> @rick-github commented on GitHub (Mar 11, 2025): Yes, image models will accept images.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#52796