Privacy & Data Boundary

Fleuron is designed for research where raw transcripts and sensitive quotes must not leave your control. This page discloses the data boundaries for both the application and this website.

1. The desktop application

Fleuron runs locally on your computer. When used offline, zero network requests occur. When you choose to use optional team collaboration features, only metadata crosses to the sync server.

Fleuron Local and Sync Data BoundaryArchitectural boundary diagram showing what data stays strictly local on this computer versus what metadata crosses to the optional sync server.THIS COMPUTER (LOCAL)Never leaves this device■ Transcript raw files and full text■ Verbatim quotes and passage text■ Original filenames and directory paths■ Analytic memos and coding memos■ Full local database (project.db)■ Interviewer names and diagnosis fields■ Auth refresh tokens and crash logsSTAYS ON LOCAL DISKHTTPS / WSSEncrypted in transitNEVERCROSSESTranscriptsQuotesMemosLocal DBPathsSYNC SERVER (OPTIONAL)Only when collaborating■ Account email and auth metadata■ De-identified study & participant labels■ Codebook structure, hierarchy & colors■ Definitions, criteria & examples■ Coded span character offsets (start/end)■ Coder display names and memberships■ Deterministic content hashes (SHA-256)ENCRYPTED SYNC METADATA

Data boundary explanation: On your computer, all transcript raw files and text, verbatim quotes, filenames, analytic and reflexive memos, and the local SQLite database remain stored locally and are never transmitted. When optional collaboration sync is used, only account email, de-identified study and participant labels, codebook structure and definitions, coded span character offsets, and study membership metadata are synced over encrypted HTTPS/WSS.

Data boundary field disclosure

CategorySynced when collaboration is activeKept strictly on local disk
Account & AccessAccount email, auth metadata, study membership, coder display name, device sync stateSession refresh tokens, local application preferences, working folder paths
Study IdentityStudy title, de-identified participant IDs, interview IDs, segment counts, content hashesRaw transcript files, full transcript text, original filenames, interview dates, diagnosis fields
CodebookCode IDs, code names, definitions, inclusion criteria, exclusion criteria, examples, hierarchy, colorNone (codebook text syncs word-for-word so teams share definitions; author them de-identified)
CodingCoder attribution, segment IDs, code IDs, character start/end offsets, timestampsVerbatim quote text, coding memos, analytic memos, reflective notes
Output & DiagnosticsNothing automaticallyExported reports, local SQLite databases, backup files, crash logs
Read docs/PRIVACY-AND-PERMISSIONS.md on GitHub →Read docs/IT-DEPLOYMENT.md on GitHub →

2. This website (fleuron.study)

The website you are reading right now collects nothing and tracks nothing. It is built as static HTML and hosted on Cloudflare Pages.

  • No analytics: We run no Google Analytics, Plausible, PostHog, or tracking pixels.
  • No cookies: No cookie banner is needed because no cookies or trackers are set.
  • No forms or inputs: We collect no emails, names, newsletters, or visitor submissions.
  • Zero third-party requests: All fonts, stylesheets, scripts, and images are self-hosted. Your browser makes zero requests to third-party CDNs.
  • Enforced by build gates: This claim is not just an intention. Every build must pass an automated CI test (npm run check:no-external) that scans every generated HTML and CSS file and fails if any third-party external request exists.