FilezConverter

WHY FILEZCONVERTER

Native editing per format, polished exports in seconds

Snow • Private RAG

Meet Snow: private RAG, open embeddings

Snow answers only from your own documents and cites the file and section behind every answer. Embeddings run on the open-source all-MiniLM-L6-V2 model, locally on the server, so indexing is free and your text is never shipped to a third party for search. Chat itself plugs into OpenAI, Anthropic, Gemini, or OpenRouter wrappers, with your own keys encrypted at rest. Bring a key for frontier models or keep it local. Either way there is no lock-in.

Why convert files here?

Drafting in Markdown, editing in Word, modelling in spreadsheets, and delivering in PDF all pull in different directions, yet the final delivery still has to look intentional. PDF ensures consistent typography, page breaks, and professional presentation across devices and print. Our four studios keep the editing experience native to each format while giving you a polished export in seconds.

How to use the studios

  1. Paste or upload: drop a file in, or paste content.
  2. Review live: edit cells, merge/split, verify spacing.
  3. Download: pick a target, share or print.

Layout-aware rendering

Each studio rebuilds your document in a standards-compliant environment, then renders a vector-based export. Headings, lists, tables, code blocks, and blockquotes keep proper spacing and hierarchy.

How your content is handled

Privacy-first

To run a conversion, your file is processed in an isolated session only for the duration of the job. It is not persistently stored, never used for any other purpose, and the session directory is purged after every response, with a manual purge control and a 30-minute sweeper as backstop.

Typography that matches

Headings scale gracefully, inline code keeps its monospace style, tables stay readable. The result looks intentional instead of improvised.

Power features

Real OOXML tables in Markdown-to-Word exports, PDF merge and split, spreadsheet cell editing, sanitized live previews, and session snapshots that survive a refresh.

Frequently asked questions

Only when a conversion explicitly needs the server, for example LibreOffice or PDF rendering lanes. Every server conversion runs in a per-session tmp directory that is deleted after the response, plus a manual purge and a 30-minute sweeper. Nothing is ever retained.

No. All four studios work without an account. Sign-in unlocks the Snow document assistant (RAG chat), AI setup, and admin tools on self-hosted instances.

Each self-hosted tenant brings their own provider keys (BYOK): chat and embedding spend belongs to the key owner, never to the operator. Local embeddings are free and never leave your infrastructure.

Markdown, Word (.docx plus legacy .doc, .odt, .rtf), spreadsheets (.xlsx, .xls, .csv, .ods), PDF (merge, split, text extract), TXT, and HTML, across the Markdown, Word, Spreadsheet, and PDF studios.

Uploads are capped at 25MB per file and Markdown inputs at 1M characters, so conversions stay fast and abuse stays out. Larger jobs should be split before upload.

Use the floating button at the bottom-right and pick Support. Bug reports, improvements, and feature requests go straight to the team, no account needed.