[release/9.8] Add support for text-to-image #6711
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
* Add ITextToImageClient * Remove URI based edit since it's not available * Add filename for edit * Add OpenAI implmentation of ITextToImageClient * Fix tests * Add tests for TextToImage * Add DeletgatingTextToImageClient and tests * Add integration test and fix some bugs * Add remaining support to MEAI for TextToImage * Make all TextToImageOptions optional These are all nullable now so that the client can use defaults where appropriate. Remove quality default since it's not consistent across models. Also remove setting ResponseFormat since this is not supported by gpt-image-1. * Address feedback * Document some exceptions * Address feedback * Make EditImageAsync plural OpenAI's image API supports multiple images and this does seem to be common functionality and a better generalization. The client library doesn't expose this yet, but we should account for it. Image models may be capable of things like "Combine the subjects of these images into a single image" or "Create a single image that uses the subject from the first image and background for the second" etc. * Address feedback and add/fix tests. * Fix bad merge * Address feedback * Fix test * Use DataContent.Name for filename. * Add extensions for EditImageAsync Extension that accepts a single DataContent and one that accepts a byte[]. I've left out streams and file paths, since these require more opinions about how to load them. I filed dotnet#6683 to address streams. * Fix test * Remove use of `_model` field. * Rename ImageToText to Image * Rename TextToImage directories to Image * Rename files TextToImage -> Image * Add new request and response type * Make GenerateImagesAsync accept ImageRequest * Remove EditImageAsync * Adding GenerateStreamingImagesAsync * Update docs * Rename ImageClient ImageGenerator * Fix up some text-to-image references * Rename Image(Options|Request|Response) * Remove `Images` from `GenerateImagesAsync` * Remove streaming method We don't yet have any good public support for streaming to vet this API We can guess at how it might behave for OpenAI, but that doesn't really give enough confidence to build the API around it. * Address feedback * Provide OpenAI an appropriate filename * Remove Style from ImageGenerationOptions
jeffhandley
approved these changes
Aug 12, 2025
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Backport Add support for text-to-image (#6648)
Experimental only API to include in 9.8 release.