What's new

First public release

The full context workbench, live: drop a document, read the Markdown next to the original, and judge the conversion by its fidelity report.

• 15 input formats — notebooks, Word, slides, spreadsheets, CSV, PDF, webpages, JSON, email, LaTeX, subtitles, Markdown — parsed entirely in your browser. Nothing is uploaded, and the tool keeps working offline.
• Five output presets — Standard, Chat, RAG, Obsidian and Archive — retarget the same parse without re-converting.
• A fidelity report, not a promise — element counts the parser detected, explicit warnings for anything lost, and a weighted QC score on every conversion.
• Real token numbers — exact o200k counts and a savings strip, computed off-thread so the app never freezes.
• Figures that actually render — extracted images become relative figures/ links that work in Obsidian, GitHub and VS Code; the .zip download bundles them to match.
• Original-document viewers — Word, Excel, slides, PDFs and notebooks render in the left panel for side-by-side checking, plus opt-in local OCR for scanned PDFs.
• Context Lab — 27 articles, 13 format guides, a field manual and before/after examples on feeding documents to LLMs.