Verifiable Architecture

Security Architecture & Verification

Do not take our privacy claim on faith. Follow this 3-step technical inspection to confirm in 30 seconds that TxtTo transmits 0 bytes of your documents or text.

Quick Answer

Verify TxtTo's zero-upload architecture in three steps: watch 0 outbound requests in DevTools Network, run conversions in Airplane Mode after loading, and audit the AES-256-GCM .redactkey local vault.

1. Three-Step DevTools Network 0-Byte Audit

Open any TxtTo tool (such as JSON Formatter, PDF to Markdown, or Anonymize Text for AI) and perform this check:

  1. Press F12 (Windows/Linux) or Cmd+Opt+I (macOS) and switch to the Network tab.
  2. Optionally check Disable cache / Filter by Fetch/XHR, then paste text or drop a sensitive PDF/CSV file.
  3. Click Run / Convert and verify the Network log shows 0 requests and 0 B transferred to any server.

2. Airplane Mode & Offline Execution Test

After opening a tool page (or installing the TxtTo PWA), disconnect Wi-Fi or turn on Airplane Mode. Formatting, XLSX/PDF parsing, GBK/Shift-JIS decoding, and PII masking continue functioning at native CPU speed because the entire engine bundle lives locally in browser memory and Cache Storage.

3. Web Crypto AES-256-GCM .redactkey Vault Specification

When using TxtTo's 8 AI PII redaction tools, entity mappings (such as [[P-01]] -> real name) reside only in volatile JavaScript memory and are erased on tab close. If you export a session for later AI-reply restoration, encryption happens strictly inside your browser's Web Crypto subtle engine:

  • Cipher: AES-256-GCM with random 96-bit IV per export
  • Key derivation: PBKDF2-SHA-256 from your passphrase with random salt (passphrase never leaves RAM)
  • Offline round-trip: re-import .redactkey locally in the Restore tab to reverse AI answers safely