GET

/v1/schemas/llmgateway/image/v1/images/edits

Accept:

kind inputactivity imagehash ec718bca…extracted Nov 13, 2026version currentsource upstream spec ↗

field	recorded source
schema	Source linked above. upstream-spec Source trace `7f27a5bd597600ebe9680565011c8e83aea98292032571349f1013bb12029cba` { "fetchedAt": 1794546640, "extractorVersion": "2", "verifiedAt": null }

Request body

18 properties · 2 required · 0 $defs

property	type	description	constraints
images*	object[]	Input image references to edit. Provide image_url as HTTPS URL or data URL.	items ≥ 1 · items ≤ 16
prompt*	string	A text description of the desired image edit.	length ≥ 1
background	enum	Background behavior for generated image output. OpenAI and Azure GPT Image models apply it natively and reject it when unsupported; `transparent` requires `png` or `webp` output. Other providers only receive it as a prompt hint.	transparent \| opaque \| auto
input_fidelity	enum	Controls fidelity to the original input image(s).	high \| low
model	string	The model to use for image editing.
n	integer	The number of edited images to generate.	1 ≤ n ≤ 10
frame_count	integer		2 ≤ n ≤ 16
frame_duration_ms	integer		20 ≤ n ≤ 2000
markup	object
output_compression	integer	Compression level for jpeg or webp output.	0 ≤ n ≤ 100
output_format	enum	Output image format. `svg` returns a base64-encoded SVG document in `b64_json`; it requires a model with SVG output (GPT Image 2.6) and `n` of 1, and does not accept `output_compression`.	png \| jpeg \| webp \| svg \| gif
svg	object	SVG author model and reasoning settings. Required when the OpenAI account has SVG model selection enabled; otherwise omit it.
quality	enum	Output quality for image models.	low \| medium \| high \| xhigh \| max \| auto
size	string	Requested output image size. Supported values depend on the model and provider.
aspect_ratio	string	The aspect ratio of the edited images (e.g. '1:1', '16:9', '4:3', '5:4'). Takes precedence over size-derived defaults.
complexity	enum	Image complexity for supported models. When omitted, the provider default applies.	low \| medium \| high
moderation	enum	Content moderation strictness for models that support it (GPT Image). 'auto' applies the default filtering, 'low' is less restrictive. Ignored by models without a moderation control.	auto \| low
service_tier	enum	Processing tier for the request, forwarded to the underlying chat completion. `flex` and `priority` are only accepted for provider/model mappings that support the tier; an unsupported tier returns a 400 `unsupported_service_tier` error.	auto \| default \| flex \| priority

Validate a payload

before you spend tokens

shell

$ curl -X POST https://modelschemas.com/v1/validate \
    -d '{"provider":"llmgateway","endpointId":"v1/images/edits","payload":{…}}'
→ {"valid": false, "errors": [{"path": "#", "keyword": "required", …}]}