mirror of
https://github.com/open-webui/open-webui.git
synced 2026-03-12 10:04:14 -05:00
enhancement: better image generation UX (e.g. semantic routing) #840
No Assignees
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: github-starred/open-webui#840
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 @colaber2 on GitHub (May 7, 2024).
Originally assigned to: @tjbck on GitHub.
The current method of generating images doesn't feel natural.
My suggestions would be (either/or):
Thanks for your efforts and for the amazing work!!
@justinh-rahb commented on GitHub (May 7, 2024):
Thanks for your ideas.. agreed 100%, I too have felt it might be more appropriate to have image generation be a part of our "Playground" interface, which itself is in need of expansion (please see #2000). We're open to suggestions and especially PRs!
@tjbck commented on GitHub (May 7, 2024):
This is actively being worked on! I appreciate your patience!
@tjbck commented on GitHub (May 7, 2024):
Related: #2004
@AdaptiveStep commented on GitHub (May 14, 2024):
I wish we could just "fork a conversation" . Like you do in LLM Studio.
@colaber2 commented on GitHub (May 16, 2024):
Noted! I will check it.
@colaber2 commented on GitHub (May 16, 2024):
Great to know!! can't wait to use it.
@francis2tm commented on GitHub (Nov 12, 2024):
@tjbck hello! Is there a PR for this already?
Cheers
@edurenye commented on GitHub (Jan 12, 2025):
In order for it to be coherent with the rest of the options, we should leave the default option as it is, that it generates the prompt and from that prompt generate an image clicking at the image icon, but have an option inside the '+' like we have for the Web Search, so you can opt in to directly answer with an image. I think that would make sense and be easy and understandable by the users.
Also, another improvement that it could have is that right now when the image has been generated, and you click again the "Generate image" button, it will regenerate the image using the same seed as before, it does not make sense, it should generate a new image, not the same image. Now the only way is to click regenerate which will generate a new prompt, and then you can generate a new image, but if you like the prompt and want to keep it, there is no way to accomplish that.
@tjbck commented on GitHub (Jan 16, 2025):
Image generation toggle has been added in dev!
@robotmikhro commented on GitHub (Feb 3, 2025):
What is actually the response format expected by open web ui? Why am I still getting an error?
@robotmikhro commented on GitHub (Feb 3, 2025):
my api respon format like this for Dall e 3:
{"data":[{"url":"https://oaidalleapiprodscus.blob.core.windows.net/private/org-e3rGuSN0lcL5faoQW8UjUDF2/user-yjVvoEMOgT0Ohyh9G1ARWeJL/img-tTXLY8j8iIGGcinX43nT9J1z.png?st=2025-02-03T15%3A10%3A06Z&se=2025-02-03T17%3A10%3A06Z&sp=r&sv=2024-08-04&sr=b&rscd=inline&rsct=image/png&skoid=d505667d-d6c1-4a0a-bac7-5c84a87759f8&sktid=a48cca56-e6da-484e-a814-9c849652bcb3&skt=2025-02-03T01%3A37%3A14Z&ske=2025-02-04T01%3A37%3A14Z&sks=b&skv=2024-08-04&sig=IzDoPj881TSdTq6AjXT5k11JXR%2BCxu6RH34LESIFHZA%3D"}],"prompt":"dog","created"