AI-powered tools that run entirely in your browser — no server, no API keys, no per-use fee. OCR (optical character recognition) for images and scanned PDFs uses Tesseract.js with an on-device model; the ~10 MB language model downloads on first use and is cached thereafter. Supports English, Spanish, French, German, Italian, and Portuguese. More capabilities — AI-powered background removal, image upscaling, and text summarisation — are on the roadmap for future phases, with the same client-side-first stance.
For **getting text out of images or scanned PDFs**, use OCR (listed as both PDF OCR and AI OCR — same engine). If the PDF already has a selectable text layer (most modern PDFs do), extraction is faster via a text-layer read rather than OCR — this lives in PDF-to-Word / PDF-to-Excel, which are deferred to Phase 9. For image effects like background removal or upscaling, those tools are coming in a future phase.
Does OCR send my images to a server?
No. Tesseract.js runs entirely in your browser. The language model (~10 MB) downloads from a public CDN on first use and is cached by your browser; after that, OCR works offline and no image data ever leaves your device.
How accurate is the OCR?
Depends heavily on input quality. Clean, well-lit, printed text at ≥200 DPI typically hits 95%+ character accuracy. Handwriting, skewed scans, or photos with shadows do worse. For critical accuracy, deskew and denoise the scan first.
Can I OCR a multi-page PDF?
Yes — the tool rasterises each page at 2× scale via pdfjs and recognises them sequentially, showing per-page progress. For very long PDFs, the first run takes time; subsequent runs in the same session are fast because the model is already loaded.
When will background removal / upscaling / summarisation ship?
These are the Phase 9 stretch tools. The scaffolding for heavy models is already in place; the integration work for each specific model is scheduled as a follow-up after the base 79 tools are polished and launched.
Are any of these tools uploading my files?
No. Every tool in this category runs entirely in your browser. Files you drop in are processed locally, never sent to a server, and discarded when you leave the page.