Build a multi-step image pipeline in n8n: turn one product photo into a cut-out, packshot and web size

Описание и применение

Автоматический конвейер в n8n преобразует одну фотографию продукта в три версии (прозрачный вырез, упаковка с белым фоном и веб‑размер) и сохраняет постоянные URL‑адреса в Google Sheets.

Задача

Manual conversion of a single product photo into three required variants (transparent cut‑out, white‑space packshot, web‑size) is time‑consuming and error‑prone when scaling to hundreds of images.

Результат

The pipeline processes a new product photo in under a minute, producing three variants with permanent URLs and recording them in a spreadsheet, eliminating duplicate uploads and manual effort.

Рабочий процесс

• Trigger on new file in Google Drive folder
• Download photo bytes
• Upload photo once to ImageStep (Asset upload)
• Branch 1: remove_bg operation (AI)
• Branch 2: pad operation (deterministic) with 120px white background
• Branch 3: resize operation (deterministic) to ≤1200px width
• Merge branch outputs
• Code node: create one row per variant with file name, variant label, assetId, publicUrl, width, height, jobId
• Append rows to Google Sheet

Доказательства

By hand that is a few minutes per photo, which is fine for five photos and not for five hundred.

Источник

One idea keeps it small: upload once, then refer to the image by its id.

Источник

The workflow, plus a Node.js version of the same pipeline, is MIT-licensed in the recipes folder of the public SDK repo.

Источник