https://cdn.filestackcontent.com/
crop_faces=mode:fill,width:400,buffer:200/
enhance/
rounded_corners=radius:max/
output=format:webp/
WV4Bqu8HQOWXxtTUHkwV
Chain multiple file operations in one request
A file processing pipeline built from one chained URL
Chain one or several image operations through the same API endpoint. One integration can transform, manipulate, edit, and process images without adding separate vendors for each step.
The whole chain is one URL, so the edge caches the result like any other asset. Change a number and the new variant is there immediately.
https://cdn.filestackcontent.com/
security=p:POLICY,s:SIGNATURE/
doc_detection=preprocess:true/
ocr/
y6FWFAcwTDipfjs6UUR0
When to use a URL chain versus a Filestack Workflow
Chaining in a URL is synchronous and image-shaped. Anything conditional, heavy, or automatic belongs in a Workflow.
| Chained URL | Workflow | |
|---|---|---|
| Runs | Synchronously, on request | Asynchronously, on upload |
| Result | The transformed file itself | A webhook with the outcome |
| Image operations | Yes | Yes |
| Conditional branching | No | Yes |
| Video and virus scanning | No | Yes |
| CDN cached | Yes | Result is stored |
Use the URL when the answer is a file and the work is image-shaped. Use a Workflow when the pipeline has to decide something, handle video, or run on every upload. An image processing workflow that only transforms belongs in the URL, one that branches belongs in a Workflow, and a media workflow that touches video always does.
File processing pipeline examples for documents and images
Explore five production patterns built from supported operations: three document pipelines and two image pipelines, each assembled as one request.
Document Intelligence
A photo of a document becomes extracted text, a sentiment score, and a routing decision.
Print-Ready Files
Customer artwork is checked for resolution and geometry, then rendered as a 300 DPI PDF.
Sensitive Documents
Capture, read, obscure the fields you nominate, strip metadata, store privately.
Profile Pictures
Find the face, frame it, clean it up, and deliver a rounded WebP from the CDN.
User-Generated Content
Screen, tag, caption, reformat, and strip metadata from a submission before it goes live.
The pipeline infrastructure Filestack runs for you
What you would build yourself
Patient intake forms
smart_crop=width:600,height:600/
auto_image/
nOtTu7K0TSLs7JvEKEMv
One URL, cached at the edge. Automated image processing with the automation already running.
Operations available in a Filestack processing chain
Grouped by category. All of these compose in a single URL, in the order you write them.
Image
resize, crop, rotate, enhance, sharpen, blur, pixelate, watermark, compress, upscale, border, circle, rounded_corners, collage
Intelligence
tags, sfw, caption, ocr, doc_detection, image_sentiment, text_sentiment
Layout engines, brand templates, and campaign managers belong to a creative automation platform, and that is their lane. This is the asset automation layer underneath one, the image workflow that runs whenever a file has to change shape, and the image processing automation you would otherwise keep alive around ImageMagick.
Frequently asked questions about file processing pipelines
What is a file processing pipeline?
A file processing pipeline is a sequence of file transformation steps applied to an uploaded file in order: cropping, analyzing, converting, and delivering it. With Filestack the sequence is segments of one URL, so an image processing pipeline built this way is a URL rather than a service you operate.
How many transformations can I chain?
Operations are applied left to right in the URL path and there is no fixed limit on how many you combine. Order matters: format conversion with output should come last, after the image manipulations.
Do chained transformations cost more than one transformation?
Transformation results are cached by the CDN, so a given chain is computed once and served from cache after that. Check the pricing page for how your plan counts transformations.
When should I use a Workflow instead of a URL?
Use a Workflow when the pipeline needs to branch on a result, process video, run virus scanning, or run automatically on every upload rather than when a URL is requested. Workflows are asynchronous and report their outcome by webhook.
Are pipeline results cached?
Yes. A chained transformation URL is cached at the CDN edge, so repeat requests for the same chain on the same file are served without reprocessing.