Privacy without
the theatre.

Browser editor

• The parser and rules run in a Web Worker on your device.
• The editor keeps the current draft, filename, and Markdown or plain-text mode in this browser’s local storage so it can recover the draft after a reload.
• Opening and downloading files use browser file APIs. SlopSift does not upload those files.
• The browser fetches parser and runtime files from SlopSift’s Cloudflare-hosted domains. Those requests do not contain the draft.
• Clearing site data for slopsift.dev removes the recovery draft.

VS Code extension

• The extension runs the bundled parser and rules in the local VS Code extension host.
• It reads the active document to produce diagnostics and does not store drafts or findings.
• There is no analytics, telemetry, advertising, or account system.
• VS Code stores extension settings such as the minimum finding level. A model download occurs only if the bundled model cannot load and the user explicitly enables that fallback.

Chrome extension

The Chrome extension is experimental and has not been published.

• Draft text moves from the page content script to SlopSift’s own extension service worker for local analysis.
• The extension packages the parser and rules. Inference does not call a remote API.
• The extension does not store drafts, findings, page contents, or browsing history.
• There is no analytics, telemetry, advertising, or account system.
• Chrome’s local extension storage keeps only the enabled state and minimum severity.

Permissions

SlopSift uses the storage permission for those two settings. Its content script runs on HTTP and HTTPS pages so it can lint editable fields. It does not inspect a field until it receives focus and contains enough text to lint.

Changes and contact

We will publish material changes here and record them in the repository before a new extension release. Report privacy or security concerns through the private process in SECURITY.md.