IdrissaMaiga/document-photo-to-text-ai

Document Photo to Text AI

A powerful Node.js library for extracting text from various document formats using AI-powered OCR and specialized parsers. Supports multiple AI providers: Google Gemini, OpenAI, and Anthropic Claude.

Handles PDFs, images, Word documents, Excel spreadsheets, CSV files, HTML pages, YouTube videos, and 40+ formats total.

Features

• Multi-Provider AI — Choose your preferred AI provider: Gemini, OpenAI, Anthropic Claude, or bring your own
• 42 Supported Formats — PDFs, images, Word, Excel, PowerPoint, CSV, HTML, JSON, XML, YAML, code files, and more
• AI-Powered OCR — Uses AI vision models for accurate text extraction from images and scanned documents
• Works Without AI — Text-based formats (TXT, JSON, CSV, DOCX, XLSX, etc.) work without any AI provider
• YouTube Support — Extracts transcripts and metadata from YouTube videos
• Web Content — Scrapes and extracts text from any web page
• Buffer Support — Process files directly from memory buffers (great for uploads)
• Built-in Caching — Automatic caching for improved performance
• Custom Providers — Plug in any AI function as a custom provider

Installation

npm install document-photo-to-text-ai

Then install the AI provider SDK you want to use (only needed for image/OCR features):

# For Google Gemini
npm install @google/generative-ai

# For OpenAI (GPT-4o, GPT-4 Vision)
npm install openai