[ ADOBE STOCK // IMAGE RECOGNIZER ]

上传图片或文件夹，AI 批量生成 Adobe Stock 标准的标题、关键词和分类Upload images or folders, and AI will batch generate Adobe Stock standard titles, keywords, and categories.

📥 下载本地程序📥 Download local program

SYSTEM OVERVIEW // Gallery recognition panorama">> SYSTEM OVERVIEW // 图库识别全景SYSTEM OVERVIEW // Gallery recognition panorama">> SYSTEM OVERVIEW // Gallery recognition panorama">>

从「看到一张图」到「安全上架一批不重复的图」全链路自动化Full-link automation from "seeing a picture" to "safely putting a batch of non-duplicate pictures"

🔎
模块B · 生成方向Module B·Generation direction
抓图库建避重库Grab the library and build a duplicate avoidance library
生成不撞车关键词Generate non-crash keywords
→
🎨
出图 / 挑图Publish/Select pictures
按提示词产出Produce by prompt word
待识别素材Material to be identified
→
🧠
模块A · 图片识别Module A·Picture Recognition
视觉模型生成Visual model generation
标题/关键词/分类Title/Keywords/Category
→
📦
上传 Adobe StockUpload Adobe Stock
导出 CSVExport CSV
批量上架Batch listing
↻ 上架后重新抓取图库 → 更新避重库 → 下一轮继续避重（闭环）↻ Re-crawl the gallery after it is put on the shelf → Update the gravity avoidance library → Continue to avoid gravity in the next round (closed loop)
MODULE B

图片识别引擎Image recognition engine

通义千问 qwen-vl-max 视觉模型批量识别，自动压缩（≤800px），多线程并发，输出符合 Adobe Stock 标准的 Title / Keywords(≤49) / Category(1-21)，另内置 qwen-max 趋势分析生成可销售提示词。Tongyi Qianwen's qwen-vl-max visual model batch recognition, automatic compression (≤800px), multi-thread concurrency, output Title / Keywords (≤49) / Category (1-21) that comply with Adobe Stock standards, and built-in qwen-max trend analysis to generate sellable prompt words.

qwen-vl-max 多线程并发Multi-thread concurrency CSV 导出CSV export 趋势提示词Trend prompt words
MODULE A

图库避重优化器Library weight avoidance optimizer

Chrome CDP 抓取 Contributor 已上图库（不碰账号密码），建立本地避重数据库；审计候选批次的关键词重叠（前10词≥5=高风险），修订至零风险后复制到剪贴板。Chrome CDP grabs the Contributor that has been uploaded to the gallery (without touching the account and password), and establishes a local avoidance database; audits the overlapping keywords of the candidate batch (the first 10 words ≥ 5 = high risk), and then copies them to the clipboard after revising them to zero risk.

CDP 抓取CDP scraping 关键词去重Keyword removal 风险审计risk audit Obsidian 同步Obsidian sync
通义千问 VLTongyi Qianwen VL qwen-max Python / Flask Pillow Tailwind Node.js + CDP Obsidian
API Config // 配置API Config // configuration

> 请确�� API Key 有效，qwen-vl-max 模型需已开通 DashScope 服务> Please make sure that the API Key is valid and the qwen-vl-max model must have activated the DashScope service.

Trend Keywords // 热点关键词生成Trend Keywords // Hot keyword generation
Upload Images // 上传图片或文件夹Upload Images // Upload pictures or folders
▵

上传单张图片Upload a single image

JPG, PNG, WEBP
▲

上传文件夹Upload folder

自动识别所有图片Automatically recognize all pictures
▸ 待处理文件（共▸ Pending documents (total 0 个）indivual)
并发数：Number of concurrencies:
[ < VIBECODING ]