One API for every file your users upload.
Stop making your users convert files for you. Send whatever arrives — Fylane returns clean, safe, application-ready output.
no credit card · signup opens with the API
Every upload form is a confession.
“JPG or PNG only, max 5 MB” is rarely the experience anyone wanted. It is the experience the team could afford to support. The restriction is not laziness — it is the cost of parsers, decoders and security work, pushed onto the person uploading the file.
A demonstration of three uploads, each shown twice. A 7.8-megabyte HEIC photo is refused by a product-image field that accepts only JPG or PNG, then accepted by Fylane and returned as three WebP sizes. An Apple Pages résumé is refused by a job application that accepts only PDF or DOCX, then returned as a PDF, Markdown text and a page preview. A 24-megabyte scanned TIFF receipt is refused for exceeding a five-megabyte limit, then returned as a searchable PDF, OCR text and extracted totals. In every case the file is unchanged; only what stood behind the upload field is different.
Three lines.
Any file.
No format branching, no worker queue, no temporary files. One call returns previews, text and metadata together.
// npm i @fylane/node
import { Fylane } from "@fylane/node";
const fylane = new Fylane(process.env.FYLANE_API_KEY);
const job = await fylane.process(upload, { preset: "ai-ready" });200 OK
{
"id": "job_01JBQ8Z5T7WXK9MNP2RSTVA3C4",
"object": "job",
"state": "queued",
"preset": "ai-ready",
"credits_estimated": 12
}An upload field is a public entry point.
Accepting arbitrary files is not one feature. It is a list of problems that each need a different answer, and every one of them is why a team narrows the field instead.
IMG_4821.HEICThe default on every recent iPhone, and unreadable by most browsers. Needs decoding and re-encoding before it is an image anyone can see.
scan_final.TIFFA different decoder again, often multi-page, frequently enormous, usually a scan that needs OCR to be worth anything.
logo.svgNot an image — a document that can carry scripts and remote references. Serving it back unsanitised is stored cross-site scripting.
photo.jpgDeclared as an image; the leading bytes say it is an executable. Trusting the extension or the browser-supplied type is how the wrong thing gets stored and served.
poster.png80,000 by 80,000 pixels. Small on disk, tens of gigabytes once decompressed into memory. A denial of service that costs the sender nothing.
archive.zipForty megabytes that expand to several terabytes. The bill lands on whoever unpacked it.
contract.pagesNot a format most backends handle, and the user has no idea it is unusual. They will either give up or email it to support.
invoice.pdfPassword-protected, or a photograph of paper with no text layer at all. Both parse cleanly and return nothing useful.
None of these are exotic. They are an ordinary week of user uploads. Fylane answers all of them before your code ever sees the file.
Anywhere the public uploads something.
The problem is worst where the person uploading is not a technical user, cannot be asked to convert anything, and will simply give up or telephone somebody. That is most of the public sector, and a great deal of finance, legal, insurance and education.
Government and public servicesregulated
A citizen photographs their proof of address at nine in the evening, sideways, on a four-year-old phone.
What arrives
HEICJPEGPDFDOCXscansAn upright, archival PDF and searchable text — without a helpdesk call.
Tax filing and accountingregulated
A year of receipts, photographed at an angle, half of them in a car, plus a bank statement export.
What arrives
HEICJPEGTIFFPDFCSVDeskewed pages, OCR text, and totals and dates extracted as JSON.
Recruitment and hiring
An applicant attaches resume.pages. They have no idea that is unusual, and your parser has never seen one.
What arrives
PDFDOCXDOCPagesJPEGA PDF to show a human and Markdown to feed a search index.
Insurance claimsregulated
Twelve photographs of a damaged bumper, a policy PDF, and a scanned police report, all in one submission.
What arrives
HEICJPEGPDFTIFFConsistent previews, metadata stripped on request, and the policy text extracted.
Banking and onboardingregulated
A passport photographed under a desk lamp, and a utility bill someone renamed to .pdf without converting it.
What arrives
HEICJPEGPDFTIFFThe real file type — not the claimed one — plus a clean image and extracted text.
Legal and conveyancing
A 400-page evidence bundle assembled from six sources, three of which are photographs of paper.
What arrives
PDFDOCXscansimagesOne merged, paginated, searchable PDF with a preview per page.
Education and assessment
An assignment submitted as a photograph of a laptop screen, forty minutes before the deadline.
What arrives
PDFDOCXODTJPEGA normalised PDF to mark and extracted text to check.
Marketplaces and commerce
A seller lists a product from their phone. The photo is 8 MB, rotated, and in a format the browser cannot show.
What arrives
HEICJPEGPNGTIFFAn oriented, metadata-stripped set of WebP and AVIF sizes, ready for a CDN.
Several of these are regulated sectors with procurement requirements attached. Fylane holds no compliance certification today — not SOC 2, HIPAA, PCI or FedRAMP — and says so plainly on the security page. If your programme requires one of them now, Fylane is not the right choice now.
Stop building file infrastructure.
- S3 wiring
- CDN config
- sharp
- LibreOffice
- ClamAV
- an OCR engine
- Redis queues
- worker fleets
await fylane.process(file)
Sixteen services. One key.
Use one of them or all of them. Nothing here is a separate product, an add-on, or a different bill.
- 01
Upload API
Direct, multipart and resumable uploads, URL ingestion, S3 imports, and a drop-in uploader.
- 02
File Validation
Real MIME verification, extension spoof detection, size and dimension limits, corrupted-file checks.
- 03
Malware Scanning
Virus and macro inspection, executables, archive bombs, quarantine with a webhook.
- 04
Image Processing
Resize, crop, rotate, compress and convert between HEIC, JPEG, PNG, WebP, AVIF and TIFF.
- 05
Image Intelligence
Background removal, face blur, auto crop, object detection, content moderation.
- 06
Document Conversion
DOCX, XLSX, PPTX, ODT and RTF to PDF or images, with fonts and layout preserved.
- 07
PDF Engine
Merge, split, compress, watermark, rotate, redact, encrypt and produce PDF/A.
- 08
OCR
Scans and photos to searchable text across 100+ languages, with deskew and cleanup.
- 09
Structured Extraction
Tables, forms, key-value pairs and named fields returned as clean JSON.
- 10
AI-Ready Documents
Any document to Markdown, chunks, tables and images, sized for embeddings and RAG.
- 11
Archive Processing
ZIP, TAR and 7z inspection, safe extraction, entry listing and recompression.
- 12
Preview & Viewer
Browser-ready previews for PDF, Office documents, spreadsheets and images.
- 13
Storage & CDN
Signed and expiring URLs, private files, global delivery, or bring your own bucket.
- 14
Workflow Engine
Chain any of the above with conditions, retries and webhooks. No queues on your side.
- 15
Video ProcessingV2
MOV, WebM and MKV to MP4 or HLS, compression, clipping, thumbnails and poster frames.
- 16
Audio ProcessingV2
Format conversion, compression, waveform generation and audio extraction from video.
One word instead of fifteen options.
Any document in, embeddable text out.