byted-mediakit-image
volcengine/mediakit-cli
Batch image processing: resize, compress, enhance, remove backgrounds, detect text, blur faces, add watermarks, and more.
What is byted-mediakit-image?
A comprehensive image manipulation toolkit supporting single and batch processing. Use it for resizing, quality optimization, content understanding (OCR, quality assessment), privacy protection (face blurring, background removal), and basic editing (rotation, cropping, color adjustment, watermarks).
- Resize and compress images with quality control (slim-image for quality-first, compress-image for format conversion)
- Detect and extract text from images using OCR; assess image quality and aesthetics
- Remove backgrounds, blur faces, and apply mosaic/blur for privacy protection
- Crop intelligently by subject or manually; rotate, flip, and add rounded corners
- Adjust brightness, contrast, saturation; sharpen, enhance, or invert images
- Add text and image watermarks; probe metadata and image properties
How to install byted-mediakit-image
npx skills add https://github.com/volcengine/mediakit-cli --skill byted-mediakit-image- mediakit-cli installed and available in PATH
- byted-mediakit-shared skill installed (required dependency)
- Cloud API credentials configured (MEDIAKIT_SURFACE and MEDIAKIT_RUNTIME environment variables)
How to use byted-mediakit-image
- 1.Install the skill: npx skills add https://github.com/volcengine/mediakit-cli --skill byted-mediakit-image
- 2.Verify mediakit-cli is available: mediakit-cli image --help
- 3.For size reduction, read the image-size-reduction family guide to choose slim-image (quality-first) or compress-image (format/size control)
- 4.Select the appropriate tool from the 21 available image operations based on your task
- 5.Run the command with required parameters; provide image paths, quality settings, or processing options as needed
- 6.Check the reference documentation for each tool for detailed parameter and output specifications
Use cases
- Batch optimize product images for e-commerce while preserving quality
- Redact sensitive information (faces, license plates, documents) before publishing or sharing
- Generate thumbnails and multi-device variants from a single source image
- Enhance low-quality user-generated content for feeds or recommendations
- Extract text from screenshots or documents for indexing or data extraction
- Content platforms and social media teams managing user uploads
- E-commerce and product photography workflows
- Privacy and compliance teams handling image sanitization
- Image processing pipelines and batch automation
- Developers building image-heavy applications
byted-mediakit-image FAQ
Use slim-image when quality is the priority and you want to reduce file size without precise constraints. Use compress-image when you need exact quality values, size limits, format conversion, or are willing to accept lossy PNG compression.
Use face-blur-image to automatically detect and pixelate all faces, mosaic-image to blur specific rectangular regions (IDs, license plates, etc.), or remove-image-background to isolate subjects and remove context.
Yes, the skill supports batch processing. Provide multiple image paths and the tool will process them according to your parameters.
The skill handles common formats and can convert between them. Check the specific tool's reference documentation for format support and conversion options.
Use the image-ocr tool, which recognizes simplified Chinese and English text, returning text blocks with position coordinates and confidence scores.
Full instructions (SKILL.md)
Source of truth, from volcengine/mediakit-cli.
name: byted-mediakit-image version: "0.2.1" license: "MIT" description: "面向单张或批量图片的视觉处理、质量优化、内容理解与基础编辑目标，适用于图片尺寸缩放与体积治理、元信息探测、裁剪旋转翻转与圆角、颜色与锐化清晰度调整、负片、模糊与打码、水印、背景移除、文字识别、画质评估与智能裁剪等。若对象和目标族已明确属于图片优化、图片理解或图片隐私保护，但具体做法不确定，可先加载本 Skill 探索。" permissions:
- shell metadata: requires: bins: ["mediakit-cli"] cliHelp: "mediakit-cli image --help" product: mediakit-cli/skills domain: image capability_count: 21
image MediaKit Skill
使用规则
- 先读取
../byted-mediakit-shared/SKILL.md，执行统一前置检查；该 Skill 缺失时停止并提示安装。 - 只从下表选择
image域工具；相似能力按各工具“能力描述”和参数边界区分。 - 执行前按需读取对应 reference；参数与结果说明来自同一份已审核文案，完整机器合同以当前 CLI
--schema为准。 - 缺少必填参数、鉴权环境变量或真实输入资源时，向用户索取；通用可选字段只能透传用户明确提供的值，其他可选字段可由明确意图准确确定，但不得伪造。
- 执行时设置
MEDIAKIT_SURFACE=skill，指定调用来源是 Skill。 - 执行时把
MEDIAKIT_RUNTIME设置为当前 Agent 宿主，避免 CLI 无法可靠识别父级 Agent。
图片体积治理与跨域路由
- 用户强调“尽量不掉画质”“质量优先”“高质量缩小体积”，且没有给出精确体积上限、质量值或格式转换要求时，先读取 reference/families/image-size-reduction.md，优先选择 slim-image。
- 用户明确给出 quality、max_size、output_format，要求格式转换，或明确接受 PNG 有损压缩时，按同一 family guide 选择 compress-image。
- 用户只说“压缩”“变小”且无法判断质量优先还是精确体积、质量或格式控制时，先澄清。
- 若只说有图片而未说明业务目标，应先澄清。把多张图片做成视频、给视频叠图或视频画面裁剪应路由到 editing；视频抽帧、视频理解、视频增强或视频字幕擦除应路由到 video。
工具列表
| 工具 | 说明 | 支持模式 | 命令 | 参考 |
|---|---|---|---|---|
| add-image-watermark | 为图片添加图文明水印，适用于版权标识与素材分发防盗链场景。 | Cloud | mediakit-cli image add-image-watermark | reference/add-image-watermark.md |
| adjust-image-color | 对输入图像的亮度、对比度和饱和度进行调整，支持调亮、调暗、增强对比度、减弱对比度、增强饱和度、减弱饱和度共 6 种快速调整效果。适用于素材基础优化、统一内容视觉风格、营造庄重、复古等特殊氛围等场景。 | Cloud | mediakit-cli image adjust-image-color | reference/adjust-image-color.md |
| compress-image | 有损图像压缩与格式转换工具；用户明确给出 quality、max_size、output_format、要求格式转换，或明确接受 PNG 有损压缩时选择。若用户强调尽量不掉画质、质量优先或高质量缩小体积，应优先选择 slim-image。 | Cloud | mediakit-cli image compress-image | reference/compress-image.md |
| crop-image | 对输入图像进行多模式裁剪，可执行方向裁剪、定向裁剪、自定义裁剪或内切圆裁剪，适用于多端尺寸适配、主体保留、商品图去边和指定区域截取。 | Cloud | mediakit-cli image crop-image | reference/crop-image.md |
| enhance-image | 基于图像内容理解进行智能决策，提升图片的分辨率、清晰度与色彩表现。 | Cloud | mediakit-cli image enhance-image | reference/enhance-image.md |
| erase-image | 可按不同场景控制自动检测并擦除图片中的文字或常见图标，擦除后的区域通过智能填充技术进行修复，修复后的区域与背景自然融合。 | Cloud | mediakit-cli image erase-image | reference/erase-image.md |
| evaluate-image-quality | 用于图像画质评估，对输入图片进行主客观画质和美学评分，适用于质量监控、低质图筛查、内容审核、推荐排序和训练数据清洗。 | Cloud | mediakit-cli image evaluate-image-quality | reference/evaluate-image-quality.md |
| face-blur-image | 自动检测图片中的所有人脸区域并进行马赛克处理，用于一键保护图片中的人脸隐私。支持社交平台内容审核、街景或监控画面脱敏、新闻媒体素材处理以及 AI 训练数据集脱敏等批量人脸隐私保护场景。 | Cloud | mediakit-cli image face-blur-image | reference/face-blur-image.md |
| flip-image | 支持对单张图片执行水平或竖直翻转。 | Cloud | mediakit-cli image flip-image | reference/flip-image.md |
| gaussian-blur-image | 用于图像高斯模糊；通过设定模糊强度快速对图片进行模糊处理，适用于隐私信息弱化、背景氛围化、生成预览图及封面背景等场景。 | Cloud | mediakit-cli image gaussian-blur-image | reference/gaussian-blur-image.md |
| image-ocr | 用于通用印刷体文字识别（OCR），识别图片中的简体中文和英文，并提供文本块位置坐标与置信度参考。 | Cloud | mediakit-cli image image-ocr | reference/image-ocr.md |
| invert-image | 用于图像负片，对输入图像执行负片（反相）效果，将图像的明暗关系与颜色映射为原图的相反效果，即明暗反转、色彩转为补色。 | Cloud | mediakit-cli image invert-image | reference/invert-image.md |
| mosaic-image | 支持对整张图像或指定矩形区域进行马赛克打码，可调整像素格形状与大小。支持用于遮挡人脸、证件信息、车牌、聊天记录等敏感内容。 | Cloud | mediakit-cli image mosaic-image | reference/mosaic-image.md |
| probe-image-metadata | 支持查询 metadata、avghue、alpha、blurhash 四种图像信息。 | Cloud | mediakit-cli image probe-image-metadata | reference/probe-image-metadata.md |
| remove-image-background | 自动识别并保留图像主体，移除背景后生成背景透明的图片，用于图像背景移除（抠图）。 | Cloud | mediakit-cli image remove-image-background | reference/remove-image-background.md |
| resize-image | 用于图像缩放，支持按指定宽高精确缩放，也可按长边、短边或等比模式缩放，适用于多端素材适配、封面与缩略图生成及批量图片预处理。 | Cloud | mediakit-cli image resize-image | reference/resize-image.md |
| rotate-image | 通过设置旋转角度和旋转背景样式对图片进行旋转处理，适用于图片方向校正、创意编辑和批量图像处理。 | Cloud | mediakit-cli image rotate-image | reference/rotate-image.md |
| round-corner-image | 为图片四角快速添加正圆或椭圆圆角，适用于头像、卡片、电商主图等常见视觉编辑场景。 | Cloud | mediakit-cli image round-corner-image | reference/round-corner-image.md |
| sharpen-image | 用于图像锐化，通过对输入图像进行锐化处理，有效增强图像的边缘细节与整体清晰度。适用于电商素材优化、UGC 画质增强、封面海报二创等场景。 | Cloud | mediakit-cli image sharpen-image | reference/sharpen-image.md |
| slim-image | 质量优先的图片瘦身工具；用户强调尽量不掉画质、质量优先或高质量缩小体积，且未要求精确体积上限、质量值或格式转换时优先选择。 | Cloud | mediakit-cli image slim-image | reference/slim-image.md |
| smart-crop-image | 自动识别图像中的主体人脸区域，并适配指定尺寸进行裁剪；支持普通人脸和动漫人脸场景。未识别到人脸时，可按预设的降级策略输出结果。 | Cloud | mediakit-cli image smart-crop-image | reference/smart-crop-image.md |
Related skills
More from volcengine/mediakit-cli and the wider catalog.
byted-mediakit-shared
Professional media processing toolkit for video editing, audio processing, image handling, and video analysis.
byted-mediakit-video
Intelligent video processing: enhancement, analysis, transcoding, subtitle extraction, privacy protection, and content generation.
byted-mediakit-audio
Process audio files and tracks: voice detection, metadata probing, transcoding, and voice-background separation.
byted-mediakit-editing
Edit audio, video, and images with timeline trimming, filters, effects, transitions, and compositing.
openviking
Activate when the user asks about any repository listed in the system prompt under 'OpenViking — Indexed Code Repositories', or when they ask about an external library, framework, or project that may have been indexed. Also activate when the user wants to add, remove, or manage repos. Always search the local codebase first before using this skill.
byted-interactai-guide
Guide for Volcano Engine InteractAI voice agent setup, configuration, and troubleshooting.