← Back

Your data stays yours.

What happens when you upload

Your TikTok export file is parsed entirely in your browser using a Web Worker. The raw JSON never leaves your device. We extract only aggregated statistics and send those to our server for AI analysis.

What we read from your export

  • Watch history (timestamps and video counts only)
  • Liked videos (counts and ratios)
  • Favorited videos (counts)
  • Search terms (aggregated by frequency)
  • Share history (method breakdown only)
  • Following list (grouped into categories, not individual usernames)

What we never read

  • Direct messages
  • Login history, IP addresses
  • Profile info, settings
  • Individual video URLs
  • Individual usernames you follow

How the AI analysis works

Creator DNA is local-only by design. The analysis runs entirely in your browser using Chrome's on-device AI (Gemini Nano). Nothing is sent anywhere — not even an aggregated summary. If your browser doesn't support on-device AI, the app will tell you and stop, rather than silently fall back to a remote server. For reference, the local model sees a small aggregated summary (~2KB) that never leaves your device:

{
  searchClusters: [
    { term: "cecred", count: 12, firstSeen: "...", lastSeen: "..." }
  ],
  stats: {
    videosWatched: 22429,
    videosLiked: 6000,
    videosFavorited: 3585,
    videosShared: 941,
    searchesCount: 1311,
    accountsFollowed: 2349
  },
  hourlyDistribution: { 0: 342, 1: 289, ... },
  dayOfWeekDistribution: { Monday: 3100, ... },
  likeToWatchRatio: 0.267,
  favoriteToLikeRatio: 0.597,
  shareMethodBreakdown: { "Copy Link": 412, ... },
  creatorCategories: [
    { category: "hair-care", count: 23, sampleUsernames: [] }
  ]
}

Data retention

Nothing is stored. No accounts, no cookies, no tracking. The aggregated summary exists in server memory only during the analysis call and is discarded immediately after. We don't even know you were here.