LLM Proxy

Appearance

Edit image ​

Edit (or extend) an existing image with a text prompt — JSON body referencing uploaded files, or multipart with raw image files.

POST/v1/images/edits

JSON or multipart. Unrecognized fields are forwarded to the provider as extension parameters rather than rejected — see Vendor extension fields.

Request — multipart
bash
curl http://localhost:8080/v1/images/edits \
  -H "Authorization: Bearer $LLM_PROXY_KEY" \
  -F image=@lighthouse.png \
  -F mask=@mask.png \
  -F model=gpt-image-1 \
  -F prompt="Replace the sky with a starry night" \
  -F size=1024x1024
Request — OpenAI SDK
python
from openai import OpenAI

client = OpenAI(base_url="http://localhost:8080/v1", api_key="sk-your-proxy-key")
img = client.images.edit(
    model="gpt-image-1",
    image=open("lighthouse.png", "rb"),
    mask=open("mask.png", "rb"),
    prompt="Replace the sky with a starry night",
)
print(img.data[0].url)
Response
json
{
  "created": 1718047200,
  "data": [
    {"url": "https://…/lighthouse-edited.png"}
  ]
}

JSON body ​

images (1–16 entries with file_id or image_url), optional mask (file_id/image_url), plus prompt, model, n, size, quality, response_format, user, background, input_fidelity, moderation, output_compression, output_format, partial_images, stream. Unknown fields reach the provider as extension parameters (see Vendor extension fields); for qwen/qwen-intl those include watermark, seed, bbox_list, color_palette and enable_sequential.

Multipart ​

File parts image and image[] (up to 16) plus a mask file; other fields as form strings. Every image must be an uploaded file, or the request fails with Multipart image edits require every image to be an uploaded file.

Response ​

Same shape as Create image: {"created", "data":[{"url"|"b64_json"}], …} with optional usage. Streamed edits emit image_edit.partial_image / image_edit.completed SSE events.

Provider support ​

openai, the openai-compatible family, gemini, and qwen providers — see the capability matrix.

Request — multipart
bash
curl http://localhost:8080/v1/images/edits \
  -H "Authorization: Bearer $LLM_PROXY_KEY" \
  -F image=@lighthouse.png \
  -F mask=@mask.png \
  -F model=gpt-image-1 \
  -F prompt="Replace the sky with a starry night" \
  -F size=1024x1024
Request — OpenAI SDK
python
from openai import OpenAI

client = OpenAI(base_url="http://localhost:8080/v1", api_key="sk-your-proxy-key")
img = client.images.edit(
    model="gpt-image-1",
    image=open("lighthouse.png", "rb"),
    mask=open("mask.png", "rb"),
    prompt="Replace the sky with a starry night",
)
print(img.data[0].url)
Response
json
{
  "created": 1718047200,
  "data": [
    {"url": "https://…/lighthouse-edited.png"}
  ]
}

MIT licensed

LLM Proxy