> ## Documentation Index > Fetch the complete documentation index at: https://apidoc.cometapi.com/llms.txt > Use this file to discover all available pages before exploring further. # Tạo chỉnh sửa hình ảnh > Sử dụng POST /v1/images/edits trong CometAPI để chỉnh sửa hình ảnh bằng tải lên multipart, mask, các model hình ảnh GPT và các tùy chọn điều khiển đầu ra hình ảnh được mã hóa. Sử dụng route này để chỉnh sửa các hình ảnh hiện có bằng tải lên multipart tương thích OpenAI trên CometAPI. ## Sử dụng route này khi * Bạn đã có một hình ảnh nguồn và muốn chỉnh sửa theo Prompt * Bạn có thể cần một mask để thay đổi có mục tiêu * Bạn có thể xử lý tải tệp multipart thay vì một yêu cầu JSON thuần ## Yêu cầu đầu tiên an toàn * Bắt đầu với một tệp PNG hoặc JPG * Bỏ qua mask cho đến khi luồng chỉnh sửa cơ bản hoạt động * Sử dụng `model: "gpt-image-2"` cho các yêu cầu chỉnh sửa hình ảnh GPT trên route này * Sử dụng một hướng dẫn ngắn yêu cầu một thay đổi dễ thấy * Đọc kết quả đã chỉnh sửa từ `data[0].b64_json` * Đặt `output_format: "jpeg"` khi bạn muốn payload JPEG * Dự kiến độ trễ lâu hơn so với tạo hình ảnh thông thường ## Hành vi của model * Các model chỉnh sửa hình ảnh GPT trên route này trả về dữ liệu hình ảnh base64 inline * `output_format` điều khiển loại hình ảnh được mã hóa bên trong `b64_json` * `response_format` chỉ quan trọng khi một model hỗ trợ đầu ra URL * `qwen-image-edit` tuân theo hành vi chỉnh sửa đặc thù của nhà cung cấp phía sau cùng một route CometAPI ## OpenAPI ````yaml api/openapi/image/openai/post-image-editing.openapi.json POST /v1/images/edits openapi: 3.1.0 info: title: Image Editing API version: 1.0.0 description: >- Edit existing images through the OpenAI-compatible CometAPI image edits route. GPT image edit models return inline base64 payloads in `data[].b64_json`, and `output_format` controls the encoded image type. servers: - url: https://api.cometapi.com security: - bearerAuth: [] paths: /v1/images/edits: post: summary: Edit images description: >- Upload one or more source images, optionally include a mask, and request an edited result with a text instruction. operationId: image_editing requestBody: required: true content: multipart/form-data: schema: type: object required: - image - prompt properties: image: type: string format: binary description: >- Source image file. Start with one PNG or JPG input for the simplest flow. prompt: type: string description: Edit instruction describing the change you want. example: Add a small red ribbon to the paper boat. model: type: string description: >- The image editing model to use. Choose a supported model from the [Models page](/overview/models). default: gpt-image-2 mask: type: string format: binary description: >- Optional PNG mask. Transparent areas mark the regions to edit. The mask dimensions must match the source image exactly. 'n': type: string description: Number of edited images to return. default: '1' quality: type: string enum: - high - medium - low description: Quality setting for models that support it. response_format: type: string enum: - url - b64_json description: >- Requested response container when supported by the selected model. GPT image edit models return `data[].b64_json`; use `output_format` to choose the encoded image type. output_format: type: string description: >- Encoded image type for GPT image edit results returned in `data[].b64_json`. For example, use `jpeg` for a JPEG payload. example: jpeg size: type: string description: Requested output size when supported by the selected model. default: model: gpt-image-2 prompt: Add a small red ribbon to the paper boat. output_format: jpeg responses: '200': description: Edited image result. content: application/json: schema: type: object required: - created - data - usage properties: created: type: integer usage: type: object properties: prompt_tokens: type: integer completion_tokens: type: integer total_tokens: type: integer prompt_tokens_details: type: object properties: cached_tokens_details: type: object properties: {} completion_tokens_details: type: object properties: {} input_tokens: type: integer output_tokens: type: integer input_tokens_details: type: object properties: image_tokens: type: integer text_tokens: type: integer cached_tokens_details: type: object properties: {} claude_cache_creation_5_m_tokens: type: integer claude_cache_creation_1_h_tokens: type: integer data: type: array items: type: object properties: b64_json: type: string description: >- Base64-encoded image payload. Decode this value to get the edited image bytes. url: type: string description: >- Temporary image URL when the selected model supports URL output. revised_prompt: type: string description: Provider-rewritten prompt, when available. background: type: string description: Background mode returned by models that expose it. output_format: type: string description: Encoded image type returned by GPT image models. quality: type: string description: Quality level returned by models that expose it. size: type: string description: Output size returned by models that expose it. example: created: 1776836647 usage: prompt_tokens: 0 completion_tokens: 0 total_tokens: 981 prompt_tokens_details: cached_tokens_details: {} completion_tokens_details: {} input_tokens: 785 output_tokens: 196 input_tokens_details: image_tokens: 768 text_tokens: 17 cached_tokens_details: {} claude_cache_creation_5_m_tokens: 0 claude_cache_creation_1_h_tokens: 0 data: - b64_json: example: created: 1781075000 background: opaque output_format: png quality: low size: 1024x1024 usage: input_tokens: 784 input_tokens_details: image_tokens: 768 text_tokens: 16 output_tokens: 196 output_tokens_details: image_tokens: 196 text_tokens: 0 total_tokens: 980 data: - b64_json: iVBORw0KGgoAAAANSUhEUgAA... x-codeSamples: - lang: Shell label: Default source: | curl https://api.cometapi.com/v1/images/edits \ -H "Authorization: Bearer $COMETAPI_KEY" \ -F image=@cat.jpg \ -F prompt="Add a small red bow tie on the cat" \ -F model=gpt-image-2 \ -F quality=low - lang: Shell label: With mask source: > # The mask is a PNG whose transparent areas mark the regions to edit. # Its dimensions must match the source image exactly. curl https://api.cometapi.com/v1/images/edits \ -H "Authorization: Bearer $COMETAPI_KEY" \ -F image=@cat.jpg \ -F mask=@mask.png \ -F prompt="Replace the masked area with a red bow tie" \ -F model=gpt-image-2 \ -F quality=low - lang: Python label: Default source: | import base64 import os from openai import OpenAI client = OpenAI( base_url="https://api.cometapi.com/v1", api_key=os.environ["COMETAPI_KEY"], ) with open("cat.jpg", "rb") as image_file: result = client.images.edit( model="gpt-image-2", image=image_file, prompt="Add a small red bow tie on the cat", quality="low", ) image_bytes = base64.b64decode(result.data[0].b64_json) with open("edited.png", "wb") as f: f.write(image_bytes) - lang: Python label: With mask source: > import base64 import os from openai import OpenAI client = OpenAI( base_url="https://api.cometapi.com/v1", api_key=os.environ["COMETAPI_KEY"], ) # The mask is a PNG whose transparent areas mark the regions to edit. # Its dimensions must match the source image exactly. with open("cat.jpg", "rb") as image_file, open("mask.png", "rb") as mask_file: result = client.images.edit( model="gpt-image-2", image=image_file, mask=mask_file, prompt="Replace the masked area with a red bow tie", quality="low", ) image_bytes = base64.b64decode(result.data[0].b64_json) with open("edited.png", "wb") as f: f.write(image_bytes) - lang: JavaScript label: Default source: > import fs from "node:fs"; import OpenAI, { toFile } from "openai"; const client = new OpenAI({ baseURL: "https://api.cometapi.com/v1", apiKey: process.env.COMETAPI_KEY, }); const result = await client.images.edit({ model: "gpt-image-2", image: await toFile(fs.createReadStream("cat.jpg"), "cat.jpg", { type: "image/jpeg" }), prompt: "Add a small red bow tie on the cat", quality: "low", }); fs.writeFileSync("edited.png", Buffer.from(result.data[0].b64_json, "base64")); - lang: JavaScript label: With mask source: > import fs from "node:fs"; import OpenAI, { toFile } from "openai"; const client = new OpenAI({ baseURL: "https://api.cometapi.com/v1", apiKey: process.env.COMETAPI_KEY, }); // The mask is a PNG whose transparent areas mark the regions to edit. // Its dimensions must match the source image exactly. const result = await client.images.edit({ model: "gpt-image-2", image: await toFile(fs.createReadStream("cat.jpg"), "cat.jpg", { type: "image/jpeg" }), mask: await toFile(fs.createReadStream("mask.png"), "mask.png", { type: "image/png" }), prompt: "Replace the masked area with a red bow tie", quality: "low", }); fs.writeFileSync("edited.png", Buffer.from(result.data[0].b64_json, "base64")); components: securitySchemes: bearerAuth: type: http scheme: bearer description: Bearer token authentication. Use your CometAPI key. ````