@pdfloom/sdk

PDFLoom's PDF engine as a library. It runs in the browser and in Node 20+, on the bytes you give it. There is no server, no upload and no account.

import { organize, protect, sign } from "@pdfloom/sdk";

const merged = await organize.mergeDocuments([fileA, fileB]);            // Uint8Array in, Uint8Array out
const locked = await protect.encryptDocument(merged, { userPassword: "secret" }); // AES-256 by default

const identity = await sign.createSigningIdentity({ commonName: "Pat Example" });
const signed = await sign.signPdfDigitally(locked_or_plain, { identity }); // appended, earlier signatures survive
const report = await sign.inspectDigitalSignatures(signed);

What is in it

| Namespace | Does | |---|---| | organize | merge, split, reorder, rotate, delete, crop, resize, n-up | | annotate, stamps, forms, detectFields | annotations, watermarks and page numbers, form fields, "make fillable" | | edit, textRemoval, redact | replace content, remove text from the page content for real, flatten-based redaction | | protect, sanitize | AES-256 / AES-128 passwords and permissions, unlock, metadata clean-up | | sign | certificate signatures (RSA and ECDSA, .p12 import including legacy 3DES/RC2 files), incremental signing, RFC 3161 timestamps, signature inspection | | convert, compress, optimizeImages, createDocument | images to PDF, size reduction that keeps text, building documents from blocks | | outline, portfolio, accessibility | bookmarks, attachments, alt text | | scan | find a page in a photo, straighten it, remove shadows (pixel arrays in, pixel arrays out) |

Page ranges are 1-based; page indices are 0-based, as in the function docs.

Trusted timestamps need a time-stamp relay URL (timestampUrl): browsers cannot reach time-stamping authorities directly. PDFLoom's own relay is https://pdfloom-api-tegminestar.azurewebsites.net/api/timestamp, which is intended for PDFLoom's own app; run your own (a 100-line proxy; see apps/api/src/routes/timestamp.ts) for production use.

Rendering, OCR, AI and the Office/HTML exports depend on pdf.js, workers or the DOM and live in the PDFLoom app, not here.

Build and test

pnpm --filter @pdfloom/sdk build
pnpm --filter @pdfloom/sdk test

Publishing (needs your npm account; free)

1. Create the pdfloom organization on npmjs.com (free for public packages), or change the package name in package.json.
2. Decide the licence and set license in package.json (it is UNLICENSED until you choose, which blocks reuse).
3. npm login, then from packages/sdk: npm publish --access public.