@niicojs/text-extractor

Text extraction helpers for Node.js. Supports PDF, Word, and PowerPoint files from ArrayBuffer or Node Buffer inputs.

Install

pnpm add @niicojs/text-extractor

Usage

import { readFile } from 'node:fs/promises';

import { extractFromPdf, extractFromPowerPoint, extractFromWord, extractText } from '@niicojs/text-extractor';

const pdfBuffer = await readFile('document.pdf');
const pdfText = await extractFromPdf(pdfBuffer);

const docxBuffer = await readFile('document.docx');
const wordText = await extractFromWord(docxBuffer);

const pptxBuffer = await readFile('presentation.pptx');
const powerPointText = await extractFromPowerPoint(pptxBuffer);

const text = await extractText(pdfBuffer, 'pdf');

API

extractText(input: ArrayBuffer | Buffer, ext: 'pdf' | 'pptx' | 'docx' | 'txt')

Dispatches to the matching extractor based on ext. For txt, it decodes the input with TextDecoder.

extractFromPdf(buffer: ArrayBuffer | Buffer)

Extracts merged text from a PDF using unpdf.

extractFromWord(buffer: ArrayBuffer | Buffer)

Extracts text from a Word document using @niicojs/word.

extractFromPowerPoint(buffer: ArrayBuffer | Buffer)

Extracts text from PowerPoint slides in slide order. Text runs within the same slide are separated by newlines, and slides are separated by blank lines.

Development

Use Vite+ (vp) for project commands.

vp install
vp check
vp test
vp pack

• vp check --fix fixes formatting and lint issues.
• vp test run tests/index.test.ts runs one test file.
• vp pack --watch builds in watch mode.

Tests

The test suite covers the public extractors and verifies that sliced Buffer inputs pass only their visible bytes to the underlying parsers.