> ## Documentation Index > Fetch the complete documentation index at: https://docs.haitoken.ai/llms.txt > Use this file to discover all available pages before exploring further. # OpenAI image edit > Edit existing images with the OpenAI Images protocol, supporting inpainting and outpainting This endpoint is compatible with the OpenAI Images protocol and generates a new image from the source images you provide and an edit prompt. It supports inpainting, outpainting, and multi-image fusion scenarios. ## Features * References source images through URL addresses or file IDs * Supports inpainting with the `mask` image to specify the edit area * Supports outpainting with the `background` image for canvas extension * Supports `input_fidelity` to control input image fidelity (`high` / `low`) * Supports `png`, `jpeg`, and `webp` output formats with `output_compression` control * Supports both `url` and `b64_json` response formats * Supports streaming (SSE) responses with intermediate results pushed via `partial_images` ## Authentication Include the `Authorization` header in the format `Bearer YOUR_API_KEY`. ## Supported image models Available image models in the Model Square, please refer to [Model List](https://portal.haitoken.ai/en/models). ## Quick example ```python Python theme={null} import requests url = "https://api.haitoken.ai/v1/images/edits" headers = { "Authorization": "Bearer YOUR_API_KEY", "Content-Type": "application/json" } data = { "model": "gpt-image-2", "images": [ {"image_url": "https://example.com/images/cat.png"} ], "prompt": "Replace the background with a beach at sunset, keeping the cat unchanged", "size": "1024x1024", "quality": "high", "n": 1 } response = requests.post(url, json=data, headers=headers) print(response.json()) ``` ```javascript Node.js theme={null} const response = await fetch('https://api.haitoken.ai/v1/images/edits', { method: 'POST', headers: { 'Authorization': 'Bearer YOUR_API_KEY', 'Content-Type': 'application/json' }, body: JSON.stringify({ model: 'gpt-image-2', images: [ { image_url: 'https://example.com/images/cat.png' } ], prompt: 'Replace the background with a beach at sunset, keeping the cat unchanged', size: '1024x1024', quality: 'high', n: 1 }) }); const data = await response.json(); console.log(data); ``` ```curl cURL theme={null} curl -X POST 'https://api.haitoken.ai/v1/images/edits' \ -H 'Authorization: Bearer YOUR_API_KEY' \ -H 'Content-Type: application/json' \ -d '{ "model": "gpt-image-2", "images": [ {"image_url": "https://example.com/images/cat.png"} ], "prompt": "Replace the background with a beach at sunset, keeping the cat unchanged", "size": "1024x1024", "quality": "high", "n": 1 }' ``` A successful edit returns the image data: ```json theme={null} { "created": 1787875200, "data": [ { "url": "https://example.com/images/edited-1.png" } ], "size": "1024x1024", "quality": "high", "usage": { "input_tokens": 3280, "output_tokens": 4160, "total_tokens": 7440, "input_tokens_details": { "image_tokens": 3268, "text_tokens": 12 }, "output_tokens_details": { "image_tokens": 4160, "text_tokens": 0 } } } ``` ## Next steps * See [OpenAI image generation](/docs/en/api-reference/images/openai/generation) to generate images from a text prompt * See [Seedream image generation](/docs/en/api-reference/images/seedream/generation) for Seedream extension parameters * See the [model list](/docs/en/api-reference/models/list-models) for available image models ## OpenAPI ````yaml en/api-reference/images/openai/edit/openapi.json POST /v1/images/edits openapi: 3.0.1 info: title: Default Module description: '' version: 1.0.0 servers: [] security: [] tags: [] paths: /v1/images/edits: post: tags: [] summary: OpenAI image edit parameters: - name: Authorization in: header description: API Key token (Bearer sk-xxx) required: true example: 'Bearer ' schema: type: string - name: Content-Type in: header description: '' example: application/json schema: type: string requestBody: content: application/json: schema: description: '' type: object properties: images: type: array items: $ref: '#/components/schemas/ImageReference' description: |- Image reference DTO. Used to reference input images in image edit requests, supporting references via URL address or file ID. description: Source images to edit prompt: type: string description: Edit prompt describing the changes to apply to the image background: type: string description: >- Background image data (for outpainting and similar extension scenarios) input_fidelity: type: string description: 'Input image fidelity control: high, low' mask: $ref: '#/components/schemas/ImageReference' description: Mask image specifying the edit area (inpainting) model: type: string description: Model name, e.g. "dall-e-2" 'n': type: integer description: Number of images to generate output_compression: type: integer description: Output compression level output_format: type: string description: Output image format, e.g. "png", "jpeg", "webp" partial_images: type: integer description: >- Number of intermediate result images pushed per update in streaming responses response_format: type: string description: Response format, "url" or "b64_json" quality: type: string description: Image quality, e.g. "standard", "hd" size: type: string description: Image size, e.g. "1024x1024" stream: type: boolean description: Whether to use streaming (SSE) responses user: type: string description: End-user identifier for monitoring and abuse detection required: - model - prompt - images examples: {} responses: '200': description: '' content: application/json: schema: type: object properties: created: type: integer description: Image generation timestamp (Unix seconds) format: int64 output_format: type: string description: Output image format echo background: type: string description: Background image data echo quality: type: string description: Image quality echo size: type: string description: Image size echo data: type: array items: description: >- Single image data. Contains the image URL or Base64 data, the revised prompt, the image size, and more. Compatible with the OpenAI DALL-E format (b64_json / revised_prompt) and common image model response formats. type: object properties: b64_json: type: string description: >- Base64-encoded image data (returned when response_format is b64_json) revised_prompt: type: string description: Prompt revised by the model (DALL-E 3 feature) url: type: string description: Publicly accessible image URL description: Generated image data list usage: description: Token and image usage statistics type: object properties: input_tokens: type: integer description: Input token consumption output_tokens: type: integer description: Output token consumption total_tokens: type: integer description: Total token consumption input_tokens_details: $ref: '#/components/schemas/InputTokensDetails' description: Input token details output_tokens_details: $ref: '#/components/schemas/OutputTokensDetails' description: Output token details required: - output_tokens_details description: '' headers: {} deprecated: false security: [] components: schemas: ImageReference: type: object properties: image_url: type: string description: Image URL address file_id: type: string description: File ID in the file system (obtained after upload) InputTokensDetails: type: object properties: image_tokens: type: integer description: Tokens consumed by the image portion text_tokens: type: integer description: Tokens consumed by the text portion OutputTokensDetails: type: object properties: image_tokens: type: integer description: Tokens consumed by the image portion text_tokens: type: integer description: Tokens consumed by the text portion ````