Batch File & Image Processing

Workflows can process whole folders of files and images on your computer in a single run — resize a hundred photos, OCR a stack of receipts, remove backgrounds, generate app icons, extract tables, and save the results to an output folder. You build it visually, no scripting.

The pipeline

A batch image pipeline is six nodes wired in a line:

Files → Loop → Image → ImageOps / Vision → Save → Then

Working with local files needs access to your machine. Use a project folder, or pair a Circuitry Server (EServer) so the Files and Save nodes can browse and write native folders directly.

Build it step by step

1. Files — drop a Files node, click Pick and choose your source folder (e.g. a folder of photos).
2. Loop — add a Loop node and connect Files → Loop. The loop iterates over each file.
3. Image — inside the loop, add an Image node so each file is loaded as an image.
4. ImageOps or Vision — add the operation node and choose what to do (see the lists below). Connect Image → operation.
5. Save — add a Save node, pick an output folder and a naming rule (keep the source name, or rename). Connect operation → Save.
6. Then — connect the loop's completion to a Then node to run anything that should happen once, after the whole batch.

Press Run to process the folder. Each file flows through the pipeline and lands in your output folder.

ImageOps operations

Fast, local pixel operations (no AI) — powered by image libraries and OpenCV:

Geometry

• Resize — scale to a width/height or percentage
• Rotate — rotate by an angle
• Flip — horizontal or vertical
• Crop — cut to a region

Format

• Convert Format — PNG ⇄ JPEG ⇄ WebP and more
• Strip Metadata — remove EXIF / location / camera data

Filter

• Edge Detection
• Watermark — overlay text or an image
• Canny Edges (OpenCV)
• Threshold (OpenCV) — binarise an image

Detect (OpenCV)

• Find Contours
• Detect Shapes

Asset Set (one image → many sizes)

• Favicon Set (Web) — the full set of favicon sizes
• App Icon Set (iOS / Android) — every app-icon size

Vision operations

AI-powered image understanding and editing — run on-device with Apple Vision (fast and free on Mac/iPad) or via cloud models for harder tasks:

Modify

• Remove Background
• Auto Enhance — exposure, contrast, sharpness
• Apply Filter — Core Image filters (sepia, mono, blur, sharpen, vibrance…)

Recognize

• Recognize Text (OCR)
• Recognize Handwriting
• Math → LaTeX — equations to LaTeX source
• Screenshot → Code — transcribe code from a screenshot
• Translate Text in Image — OCR + translate (signs, menus, foreign docs)
• Chart → Data — read a bar/line/pie chart into structured data
• Table → JSON — extract tabular data

Reason

• Describe Image — natural-language caption
• Generate Alt Text — short accessibility caption
• Visual Q&A — answer a question about the image
• Extract Structured Data — pull JSON from receipts, business cards, forms (to a schema)
• Classify (Open Labels) — pick from your own label set
• Count Objects
• Tag / Keyword — hashtags/keywords for content management
• Content Moderation — flag NSFW / violence / hate-symbol content
• Mock → HTML — sketch/wireframe to HTML
• Mock → React Component — UI mockup to React/JSX

Examples

• Bulk resize + convert — Files (folder) → Loop → Image → ImageOps (Resize, then Convert to WebP) → Save (output folder).
• Receipt OCR to data — Files → Loop → Image → Vision (Extract Structured Data) → Save (write the JSON) → Then (summarise the batch with an Agent).
• Background removal — Files → Loop → Image → Vision (Remove Background) → Save (PNG with transparency).
• App icon generator — Files (one logo) → Image → ImageOps (App Icon Set) → Save (output folder of every size).

Next steps

• ImageOps Node and Vision Node references
• Loop Node — how iteration works
• Circuitry Server — native local file access and faster processing