github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #13897] Support Image Editing (Img2Img/Inpainting) pipelines and API options #9093

New Issue
Closed
opened 2026-04-12 21:56:36 -05:00 by GiteaMirror · 4 comments
GiteaMirror commented 2026-04-12 21:56:36 -05:00
Owner
Copy Link

Originally created by @larria on GitHub (Jan 25, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/13897

FLUX family (e.g., flux2-klein), excel not just at Text-to-Image generation but also at Image-to-Image editing and structure retention. Currently, there seems to be no direct API parameter to control image editing workflows (like denoising strength/guidance) effectively.

I would like to request support for Image Editing (Img2Img) pipelines.
Specifically, please expose parameters to control the influence of the input image.

Extend the /api/generate endpoint to accept an images parameter for diffusion models, along with a control parameter (e.g., strength or denoising_strength).

Thanks for the recent updates regarding text-to-image capabilities after all!

Originally created by @larria on GitHub (Jan 25, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/13897 FLUX family (e.g., flux2-klein), excel not just at Text-to-Image generation but also at Image-to-Image editing and structure retention. Currently, there seems to be no direct API parameter to control image editing workflows (like denoising strength/guidance) effectively. I would like to request support for Image Editing (Img2Img) pipelines. Specifically, please expose parameters to control the influence of the input image. Extend the /api/generate endpoint to accept an images parameter for diffusion models, along with a control parameter (e.g., strength or denoising_strength). Thanks for the recent updates regarding text-to-image capabilities after all!
GiteaMirror added the feature request label 2026-04-12 21:56:36 -05:00
GiteaMirror commented 2026-04-12 21:56:39 -05:00
Author
Owner
Copy Link

@thefanclub commented on GitHub (Jan 26, 2026):

Ollama 0.15.1 on MacOS supports image editing (img2img) in the API. Just add the 'images' parameter to your api call with list of encoded base64 strings (2 images max at this point is seems). Works on both Flux2 [Klein] and Z-Image-Turbo. The parameters to control this is unknown at this point.

Big thank you to the Ollama team for making this possible even on a 8GB Macbook Air. Incredible technology.

<!-- gh-comment-id:3800576329 --> @thefanclub commented on GitHub (Jan 26, 2026): Ollama 0.15.1 on MacOS supports image editing (img2img) in the API. Just add the 'images' parameter to your api call with list of encoded base64 strings (2 images max at this point is seems). Works on both Flux2 [Klein] and Z-Image-Turbo. The parameters to control this is unknown at this point. Big thank you to the Ollama team for making this possible even on a 8GB Macbook Air. Incredible technology.
GiteaMirror commented 2026-04-12 21:56:40 -05:00
Author
Owner
Copy Link

@hqnghi88 commented on GitHub (Jan 27, 2026):

print(f"Input Image: {input_image_path}")

payload = {
"model": model,
"prompt": prompt,
"images": [base64_image],
"options": {
"temperature": 0 # Deterministic
}
}

try:
response = requests.post(url, json=payload)
This give me
Input Image: a.jpg
Error: 500
{"error":"model does not support image editing"}
How can it work for you?

<!-- gh-comment-id:3804651578 --> @hqnghi88 commented on GitHub (Jan 27, 2026): print(f"Input Image: {input_image_path}") payload = { "model": model, "prompt": prompt, "images": [base64_image], "options": { "temperature": 0 # Deterministic } } try: response = requests.post(url, json=payload) This give me Input Image: a.jpg Error: 500 {"error":"model does not support image editing"} How can it work for you?
GiteaMirror commented 2026-04-12 21:56:41 -05:00
Author
Owner
Copy Link

@thefanclub commented on GitHub (Jan 27, 2026):

NOTE : This is MacOS only at the moment with Ollama 0.15.2:

print(f"Input Image: {input_image_path}")

with open(input_image_path, "rb") as image_file:
    # Encode without base64 PNG headers. Ollama wants string only
    pngImageB64String = base64.b64encode(image_file.read()).decode('utf-8')

ref_images = []
ref_images.append(pngImageB64String)

url = 'http://localhost:11434/api/generate'
payload = {'model' : 'x/flux2-klein:4b', 
        'prompt' : prompt,
        'images' : ref_images,
        'stream' : False
 }

# Add headers
headers = {'Content-type': 'application/json', 'Accept': 'text/plain'}

response = requests.post(url, data=json.dumps(payload), headers=headers).json()

print (response)
<!-- gh-comment-id:3805577008 --> @thefanclub commented on GitHub (Jan 27, 2026): NOTE : This is MacOS only at the moment with Ollama 0.15.2: ``` print(f"Input Image: {input_image_path}") with open(input_image_path, "rb") as image_file: # Encode without base64 PNG headers. Ollama wants string only pngImageB64String = base64.b64encode(image_file.read()).decode('utf-8') ref_images = [] ref_images.append(pngImageB64String) url = 'http://localhost:11434/api/generate' payload = {'model' : 'x/flux2-klein:4b', 'prompt' : prompt, 'images' : ref_images, 'stream' : False } # Add headers headers = {'Content-type': 'application/json', 'Accept': 'text/plain'} response = requests.post(url, data=json.dumps(payload), headers=headers).json() print (response) ```
GiteaMirror commented 2026-04-12 21:56:43 -05:00
Author
Owner
Copy Link

@larria commented on GitHub (Feb 5, 2026):

Thx a lot bros, it works well!!!!

<!-- gh-comment-id:3851609015 --> @larria commented on GitHub (Feb 5, 2026): Thx a lot bros, it works well!!!!
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label feature request
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#9093
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?