All Tools
T
DataFreeOpen Source
TEXT-EXTRACT-API
Self-hosted document extraction API with OCR + LLM
MIT
ABOUT
Extracting clean, structured text from PDFs, scanned documents, and Office files is a persistent pain point for RAG pipelines and LLM data ingestion. Existing parsers either lose tabular structure, fail on scanned documents without OCR, or require complex multi-tool pipelines. text-extract-api solves this with a single Docker-deployable API that uses modern OCR (EasyOCR + PyTorch) and optional LLM enhancement (via Ollama) to convert any document or image into clean Markdown or structured JSON. It handles tables, math formulas, multi-column layouts, and includes built-in PII redaction — all with a simple HTTP endpoint.
INTEGRATION GUIDE
1. Deploy a self-hosted document processing pipeline that converts PDFs and scanned docs to LLM-ready Markdown
2. Extract structured JSON data from invoices, receipts, and forms using OCR + LLM correction
3. Build a document ingestion API for RAG systems that handles PDF, DOCX, PPTX, and images through one endpoint
4. Anonymize sensitive documents with built-in PII redaction before sending them to LLM pipelines
5. Process batch document conversions with async task queues (Celery + Redis) for high-throughput workloads
TAGS
datadocument-processingocrpdfextractionmarkdownjsonself-hosted