Workflow operations

Workflows are built from operations — the individual building blocks that tell PixelFiddler what to do with your images. Each operation appears as a color-coded node in the workflow editor, and you connect them together to create your processing pipeline.

This page covers every operation type, what it does, its key settings, and when you would reach for it.

---

Transform Image

Color: blue node

The Transform Image operation is the workhorse of most workflows. It resizes images, converts them to a different file format, and applies visual effects — all in a single step.

Key settings

Output format — Choose the file format for the processed image. Your options are:

• JPG — Best for photographs and complex images with many colors
• PNG — Preserves transparency; ideal for logos and graphics
• WebP — Smaller file sizes than JPG or PNG with similar quality
• AVIF — The smallest files, though not supported by every platform
• GIF — For simple animations or very low-color graphics
• AUTO — Lets PixelFiddler pick the best format automatically

Quality — A value from 1 to 100 that controls the compression level. Higher values mean better-looking images but larger files. A quality of 80 is a good starting point for most product photos.

Resize dimensions — Set a target width (up to 4096 px) and/or height (up to 4096 px). If you only set one dimension, the other scales proportionally.

Resize mode — Controls how the image fits into the target dimensions:

• FIT — Scales the image to fit inside the frame while keeping the original aspect ratio. The result may be smaller than the target on one side.
• FILL — Stretches the image to completely fill the frame. The aspect ratio may not be preserved.
• PAD — Keeps the aspect ratio and adds padding (a background color you choose) to fill any remaining space.

Effects — Apply one or more visual adjustments:

Effect	What it does
Blur	Softens the image (0 — 100)
Brightness	Lightens or darkens (0 — 100)
Contrast	Increases or reduces contrast (0 — 100)
Saturation	Boosts or mutes colors (0 — 100)
Sharpen	Sharpens fine detail (0 — 100)
Noise	Adds film-grain texture (0 — 100)
Grayscale	Converts to black and white
Sepia	Applies a warm, vintage tone
Rotate	Rotates by a specified angle in degrees
Recolor	Swaps one color for another across the image

Watermark — Overlay a previously uploaded watermark image. You can control its position, size, and opacity.

Crop — Cut the image to specific dimensions. Cropping supports automatic focus detection (center, smart, detail, or corner-based) as well as manual coordinates. You can also crop around a detected object such as a face or upper body.

Border — Add a border with a configurable stroke width, color, corner radius, and opacity.

Shadow — Apply a drop shadow behind the image with adjustable color, blur, offset, and opacity.

Grid — Overlay a semi-transparent grid pattern for anti-copy protection.

Text — Render text directly on the image with full control over font, color, size, position, and shadow.

Tip

If you only need to change the format or resize, leave the effects section untouched. The operation will skip any effect that is not explicitly configured, keeping processing fast.

When to use it

• Resize product images to meet marketplace dimension requirements (for example, 1500 x 1500 px for Amazon)
• Convert uploaded PNGs to WebP to cut file size before syncing to your shop
• Add a watermark to every image in a batch
• Apply consistent brightness and contrast adjustments to a product photo set

---

Background Removal

Color: emerald node

The Background Removal operation uses AI to cut out the background from your image, leaving only the main subject on a transparent canvas. The output is always a transparent PNG.

Key settings

AI model — Choose the model that processes the removal:

• BASIC — Faster and uses fewer AI credits. Good enough for simple, well-lit product photos on a plain background.
• QUALITY — More accurate on complex scenes, fine hair, or semi-transparent edges. Uses more AI credits per image.

If you do not select a model, BASIC is used by default.

When to use it

• Prepare clean product photos for Amazon, eBay, or Shopify that require a pure white or transparent background
• Feed the transparent result into a downstream Transform Image node to place the product on a solid color or add a shadow
• Process hundreds of supplier photos in one go without manual editing

Note

Because the output is always a transparent PNG, it pairs well with a Transform Image node placed after it. You can convert the PNG to JPG with a white background, or to WebP to save space, depending on what your platform requires.

---

AI Edit

Color: purple node

The AI Edit operation lets you modify an image by describing the change you want in plain language. Instead of adjusting sliders, you write a short prompt and the AI applies the edit.

Key settings

Prompt — A text description of the edit you want. For example: “Make the lighting warmer”, “Remove the wrinkles from the fabric”, or “Add a slight reflection under the product”.

Strength — A value from 0.0 to 1.0 that controls how aggressively the AI changes the image. A strength of 0.95 (the default) applies the edit with high confidence. Lower values produce subtler changes that stay closer to the original.

Model — Choose the AI model for the edit:

• QUALITY — Faithful to your prompt, producing clean and predictable edits. This is the default.
• CREATIVE — More interpretive; may surprise you with artistic choices. Useful for stylistic or mood-based edits.

When to use it

• Automatically touch up lighting or color cast across a batch of product photos
• Smooth out fabric wrinkles or minor blemishes without a photo editor
• Add subtle visual polish (reflections, depth, atmosphere) to lifestyle images

Caution

AI edits consume AI credits. The cost depends on the model and the resolution of the image. For high-volume workflows, consider using Transform Image effects (brightness, contrast, saturation) instead when a simple adjustment would achieve the same result.

---

File Format Router

Color: amber node

The File Format Router sends each incoming image down a different path based on its file format. Think of it as a sorting station: JPGs go one way, PNGs go another, and so on.

How it works

1. Add a File Format Router node to your workflow.
2. Connect multiple downstream nodes to it.
3. Assign a file format (JPG, PNG, WebP, AVIF, GIF, or others) to each outgoing connection.

When an image enters the router, PixelFiddler checks its format and sends it down the matching branch. Images whose format does not match any branch are skipped.

When to use it

• Apply different quality settings to JPGs and PNGs (photographs versus graphics)
• Convert only WebP uploads to JPG for platforms that do not accept WebP, while leaving JPGs untouched
• Route transparent PNGs to a background-removal path while sending JPGs straight to save

Tip

Combine the File Format Router with a Broadcast Router when you need both format-based branching and multiple outputs. For example, route PNGs through background removal first, then broadcast the result into several sizes.

---

Broadcast Router

Color: teal node

The Broadcast Router sends a copy of the image to every connected downstream path at the same time. It does not alter the image — it simply duplicates the flow so you can process the same input in multiple ways.

How it works

1. Add a Broadcast Router node to your workflow.
2. Connect two or more downstream nodes.
3. Each connected node receives an identical copy of the image.

Every branch runs independently, so you can apply completely different transformations on each path.

When to use it

• Generate multiple sizes from a single upload (thumbnail, medium, and full-size)
• Create both a WebP version for your website and a JPG version for marketplace feeds
• Produce a watermarked copy for public galleries alongside an unwatermarked archive copy

---

Save to Storage

Color: green node (terminal)

The Save to Storage operation writes the processed image to a storage location. This is always the last node in a workflow path — it cannot have outgoing connections.

Key settings

Target storage — The storage where the file will be saved. Select any storage connected to your space (Hosted, S3, Azure, GCP, or HTTP).

Filename template — Controls the name of the saved file. The file extension is determined automatically by the format of the processed image. You can use the following template variables:

Variable	Description	Example output
{orig_name}	Original file name without extension	photo
{date_utc}	Current UTC date in yyyy-MM-dd format	2026-02-27

For example, a template of {orig_name}-{date_utc} with an input file called photo.jpg that was converted to WebP would produce photo-2026-02-27.webp.

Target directory — An optional folder path within the target storage. If you leave this empty, files are saved to the root directory.

When to use it

Every workflow needs at least one Save to Storage node. Without it, processed images have nowhere to go. You can have multiple Save to Storage nodes in a single workflow — one at the end of each branch — to write different versions to different locations.

Note

Make sure the target storage has write permissions enabled. Hosted storage always supports writing. For external storage (S3, Azure, GCP), verify that the connected credentials allow upload operations.

---

Apply Template (coming soon)

Color: disabled node

The Apply Template operation will let you overlay your image onto a pre-designed template — for example, placing a product photo into a lifestyle scene or adding a branded frame around it.

This operation is not yet available in the workflow editor. It will be enabled in a future release.

---

How operations connect

Operations are linked together on the workflow canvas using connections (the lines between nodes). A few rules to keep in mind:

• Every workflow starts with a trigger (such as “file uploaded”) that feeds into the first operation.
• Most operations accept one input and produce one output, which you connect to the next node.
• Routers (File Format Router and Broadcast Router) fan out to multiple downstream nodes.
• Save to Storage is always a terminal node — nothing can follow it.
• You can chain as many operations as you need. A common pattern is: trigger, Transform Image, Save to Storage.

Tip

Start simple. A workflow with just a Transform Image node and a Save to Storage node is enough to automate resizing on upload. You can always add more operations later.

---

Next steps

Create your first workflow

Resize images on upload

Automatic background removal