Kreuzberg document intelligence - WebAssembly bindings
npm install @kreuzberg/wasm
Extract text, tables, images, and metadata from 62+ file formats including PDF, Office documents, and images. WebAssembly bindings for browsers, Deno, and Cloudflare Workers with portable deployment and multi-threading support.
Install via one of the supported package managers:
npm:
``bash`
npm install @kreuzberg/wasm
pnpm:
`bash`
pnpm add @kreuzberg/wasm
yarn:
`bash`
yarn add @kreuzberg/wasm
- Modern browser with WebAssembly support, or Deno 1.0+, or Cloudflare Workers
- Optional: Tesseract WASM for OCR functionality
Extract text, metadata, and structure from any supported document format:
`ts
import { extractBytes, initWasm } from "@kreuzberg/wasm";
async function main() {
await initWasm();
const buffer = await fetch("document.pdf").then((r) => r.arrayBuffer());
const bytes = new Uint8Array(buffer);
const result = await extractBytes(bytes, "application/pdf");
console.log("Extracted content:");
console.log(result.content);
console.log("MIME type:", result.mimeType);
console.log("Metadata:", result.metadata);
}
main().catch(console.error);
`
#### Extract with Custom Configuration
Most use cases benefit from configuration to control extraction behavior:
With OCR (for scanned documents):
`ts
import { enableOcr, extractBytes, initWasm } from "@kreuzberg/wasm";
async function extractWithOcr() {
await initWasm();
try {
await enableOcr();
console.log("OCR enabled successfully");
} catch (error) {
console.error("Failed to enable OCR:", error);
return;
}
const bytes = new Uint8Array(await fetch("scanned-page.png").then((r) => r.arrayBuffer()));
const result = await extractBytes(bytes, "image/png", {
ocr: {
backend: "tesseract-wasm",
language: "eng",
},
});
console.log("Extracted text:");
console.log(result.content);
}
extractWithOcr().catch(console.error);
`
#### Table Extraction
See Table Extraction Guide for detailed examples.
#### Processing Multiple Files
`ts
import { extractBytes, initWasm } from "@kreuzberg/wasm";
interface DocumentJob {
name: string;
bytes: Uint8Array;
mimeType: string;
}
async function _processBatch(documents: DocumentJob[], concurrency: number = 3) {
await initWasm();
const results: Record
const queue = [...documents];
const workers = Array(concurrency)
.fill(null)
.map(async () => {
while (queue.length > 0) {
const doc = queue.shift();
if (!doc) break;
try {
const result = await extractBytes(doc.bytes, doc.mimeType);
results[doc.name] = result.content;
} catch (error) {
console.error(Failed to process ${doc.name}:, error);
}
}
});
await Promise.all(workers);
return results;
}
`
#### Async Processing
For non-blocking document processing:
`ts
import { extractBytes, getWasmCapabilities, initWasm } from "@kreuzberg/wasm";
async function extractDocuments(files: Uint8Array[], mimeTypes: string[]) {
const caps = getWasmCapabilities();
if (!caps.hasWasm) {
throw new Error("WebAssembly not supported");
}
await initWasm();
const results = await Promise.all(files.map((bytes, index) => extractBytes(bytes, mimeTypes[index])));
return results.map((r) => ({
content: r.content,
pageCount: r.metadata?.pageCount,
}));
}
const fileBytes = [new Uint8Array([1, 2, 3])];
const mimes = ["application/pdf"];
extractDocuments(fileBytes, mimes)
.then((results) => console.log(results))
.catch(console.error);
`
- Installation Guide - Platform-specific setup
- API Documentation - Complete API reference
- Examples & Guides - Full code examples and usage guides
- Configuration Guide - Advanced configuration options
62+ file formats across 8 major categories with intelligent format detection and comprehensive metadata extraction.
#### Office Documents
| Category | Formats | Capabilities |
|----------|---------|--------------|
| Word Processing | .docx, .odt | Full text, tables, images, metadata, styles |.xlsx
| Spreadsheets | , .xlsm, .xlsb, .xls, .xla, .xlam, .xltm, .ods | Sheet data, formulas, cell metadata, charts |.pptx
| Presentations | , .ppt, .ppsx | Slides, speaker notes, images, metadata |.pdf
| PDF | | Text, tables, images, metadata, OCR support |.epub
| eBooks | , .fb2 | Chapters, metadata, embedded resources |
#### Images (OCR-Enabled)
| Category | Formats | Features |
|----------|---------|----------|
| Raster | .png, .jpg, .jpeg, .gif, .webp, .bmp, .tiff, .tif | OCR, table detection, EXIF metadata, dimensions, color space |.jp2
| Advanced | , .jpx, .jpm, .mj2, .jbig2, .jb2, .pnm, .pbm, .pgm, .ppm | OCR via hayro-jpeg2000 (pure Rust decoder), JBIG2 support, table detection, format-specific metadata |.svg
| Vector | | DOM parsing, embedded text, graphics metadata |
#### Web & Data
| Category | Formats | Features |
|----------|---------|----------|
| Markup | .html, .htm, .xhtml, .xml, .svg | DOM parsing, metadata (Open Graph, Twitter Card), link extraction |.json
| Structured Data | , .yaml, .yml, .toml, .csv, .tsv | Schema detection, nested structures, validation |.txt
| Text & Markdown | , .md, .markdown, .djot, .rst, .org, .rtf | CommonMark, GFM, Djot, reStructuredText, Org Mode |
#### Email & Archives
| Category | Formats | Features |
|----------|---------|----------|
| Email | .eml, .msg | Headers, body (HTML/plain), attachments, threading |.zip
| Archives | , .tar, .tgz, .gz, .7z | File listing, nested archives, metadata |
#### Academic & Scientific
| Category | Formats | Features |
|----------|---------|----------|
| Citations | .bib, .biblatex, .ris, .nbib, .enw, .csl | Structured parsing: RIS (structured), PubMed/MEDLINE, EndNote XML (structured), BibTeX, CSL JSON |.tex
| Scientific | , .latex, .typst, .jats, .ipynb, .docbook | LaTeX, Jupyter notebooks, PubMed JATS |.opml
| Documentation | , .pod, .mdoc, .troff | Technical documentation formats |
- Text Extraction - Extract all text content with position and formatting information
- Metadata Extraction - Retrieve document properties, creation date, author, etc.
- Table Extraction - Parse tables with structure and cell content preservation
- Image Extraction - Extract embedded images and render page previews
- OCR Support - Integrate multiple OCR backends for scanned documents
- Async/Await - Non-blocking document processing with concurrent operations
- Plugin System - Extensible post-processing for custom text transformation
- Batch Processing - Efficiently process multiple documents in parallel
- Memory Efficient - Stream large files without loading entirely into memory
- Language Detection - Detect and support multiple languages in documents
- Configuration - Fine-grained control over extraction behavior
| Format | Speed | Memory | Notes |
|--------|-------|--------|-------|
| PDF (text) | 10-100 MB/s | ~50MB per doc | Fastest extraction |
| Office docs | 20-200 MB/s | ~100MB per doc | DOCX, XLSX, PPTX |
| Images (OCR) | 1-5 MB/s | Variable | Depends on OCR backend |
| Archives | 5-50 MB/s | ~200MB per doc | ZIP, TAR, etc. |
| Web formats | 50-200 MB/s | Streaming | HTML, XML, JSON |
Kreuzberg supports multiple OCR backends for extracting text from scanned documents and images:
- Tesseract-Wasm
`ts
import { enableOcr, extractBytes, initWasm } from "@kreuzberg/wasm";
async function extractWithOcr() {
await initWasm();
try {
await enableOcr();
console.log("OCR enabled successfully");
} catch (error) {
console.error("Failed to enable OCR:", error);
return;
}
const bytes = new Uint8Array(await fetch("scanned-page.png").then((r) => r.arrayBuffer()));
const result = await extractBytes(bytes, "image/png", {
ocr: {
backend: "tesseract-wasm",
language: "eng",
},
});
console.log("Extracted text:");
console.log(result.content);
}
extractWithOcr().catch(console.error);
`
This binding provides full async/await support for non-blocking document processing:
`ts
import { extractBytes, getWasmCapabilities, initWasm } from "@kreuzberg/wasm";
async function extractDocuments(files: Uint8Array[], mimeTypes: string[]) {
const caps = getWasmCapabilities();
if (!caps.hasWasm) {
throw new Error("WebAssembly not supported");
}
await initWasm();
const results = await Promise.all(files.map((bytes, index) => extractBytes(bytes, mimeTypes[index])));
return results.map((r) => ({
content: r.content,
pageCount: r.metadata?.pageCount,
}));
}
const fileBytes = [new Uint8Array([1, 2, 3])];
const mimes = ["application/pdf"];
extractDocuments(fileBytes, mimes)
.then((results) => console.log(results))
.catch(console.error);
`
Kreuzberg supports extensible post-processing plugins for custom text transformation and filtering.
For detailed plugin documentation, visit Plugin System Guide.
Process multiple documents efficiently:
`ts
import { extractBytes, initWasm } from "@kreuzberg/wasm";
interface DocumentJob {
name: string;
bytes: Uint8Array;
mimeType: string;
}
async function _processBatch(documents: DocumentJob[], concurrency: number = 3) {
await initWasm();
const results: Record
const queue = [...documents];
const workers = Array(concurrency)
.fill(null)
.map(async () => {
while (queue.length > 0) {
const doc = queue.shift();
if (!doc) break;
try {
const result = await extractBytes(doc.bytes, doc.mimeType);
results[doc.name] = result.content;
} catch (error) {
console.error(Failed to process ${doc.name}:, error);
}
}
});
await Promise.all(workers);
return results;
}
``
For advanced configuration options including language detection, table extraction, OCR settings, and more:
- Official Documentation
- API Reference
- Examples & Guides
Contributions are welcome! See Contributing Guide.
MIT License - see LICENSE file for details.
- Discord Community: Join our Discord
- GitHub Issues: Report bugs
- Discussions: Ask questions