Licenses & practical limits
Your files stay on your device
Inputs, decoded data and results use browser memory. Nothing is uploaded. Results are lost when this tab closes. Limits: 50 files; 80 MB per input; 180 MB input batch; 300 MB retained results. Archives are limited to 1,000 entries and 180 MB expanded data. PDF work is limited to 200 selected pages; OCR to 20 pages. Large decoded files can exhaust memory below those limits.
What conversions preserve
Image conversion uses 8-bit sRGB and removes metadata. TIFF uses a selected page; HEIC its first image. Image conversion flattens animation. Select a video output for GIF animation. SVG export traces pixels into new paths. RAW output develops the camera data to 8-bit sRGB; it does not preserve original sensor data or every camera variant.
Audio/video use their first streams. Metadata copying is optional; chapters and subtitles are removed. Lossy formats cannot regain detail. Browser preview support differs from processor support. HDR and professional color workflows are outside this version.
Document conversion preserves text and simple heading structure, rather than complex page layout, embedded media, annotations or tracked changes. PDF-to-DOCX exports embedded text; English OCR is a separate operation for scanned pages. OCR is approximate and needs review. The bundled PDF font supports many Latin, Greek and Cyrillic characters; unsupported characters trigger a clear error rather than silent replacement.
Workbook conversion does not recalculate formulas or preserve macros/charts. Flat table formats use the selected sheet. CSV protection prefixes formula-like strings by default; disable it in Settings when exact text interoperability is required. Typed XML preserves JSON values through an explicit schema. Arbitrary nested data needs flattening rules before becoming a table.
Encrypted ZIP, ZIP64, encrypted entries, archive links and RAR/7Z export are unavailable. RAR/7Z/TAR.BZ2/TAR.XZ import is supported; standalone BZIP2/XZ streams are not enabled. Password-protected PDFs support password-based rendering/text extraction; page editing needs a decrypted copy. Full Office layout reconstruction, legacy DOC/RTF/PPT, additional OCR languages, and RAW export remain future work.
Open-source credits
Full dependency inventory, source links and license notices
FFmpeg notices · Image codecs and HEIC sources · LibRaw license · LibRaw source · RAW build recipe