PDF Text Extractor

**Quick judgement** A strong, well-scoped PDF text and OCR extraction skill suitable for most document-ingestion and digitization workflows. It offers good coverage of common needs (text vs. scanned PDFs, batch processing, multiple output formats, basic analytics) with no external dependencies, making it attractive for local or air‑gapped setups. **Strengths** - Works on both text-based and image-only PDFs using PDF.js + Tesseract.js under the hood. - Zero external dependencies and bundled engines simplify setup and improve reproducibility. - Batch processing, progress tracking, and aggregate stats fit well into automated pipelines. - Useful extras (word/char counts, language detection, metadata extraction) reduce the need for additional tooling. - Configurable OCR quality/speed and language options for trade-offs between performance and accuracy. **Key risks / limitations** - OCR accuracy (85–95% claimed) will still depend heavily on scan quality; noisy or low-DPI documents may yield poor results. - Tesseract.js OCR is CPU- and memory-intensive; large batches or very long PDFs could be slow or resource-heavy in constrained environments. - Table and structured data extraction are limited to post-processing of raw text; advanced table/field extraction is on the roadmap, not fully realized. - Minor API inconsistencies in the docs (object vs positional params in examples) may lead to some integration friction until verified against actual code. **Recommended scenarios** - Local/offline or zero-dependency environments where installing native PDF/OCR stacks is undesirable or impossible. - Ingestion pipelines for invoices, contracts, and reports where you primarily need reasonably accurate text plus basic metadata. - Preprocessing PDFs for LLMs, search indexing, or analytics, especially when you can tolerate typical OCR imperfections on scanned pages. **Less ideal for** - Workloads requiring highly accurate structured extraction (e.g., robust table recognition, precise field extraction) out of the box. - Massive-scale OCR processing where native, highly optimized OCR stacks or GPU-accelerated solutions are required for throughput and cost efficiency.