Incognito Prompt

Strips sensitive data (file paths, project names, PII) from prompts before they reach LLMs.

Incognito Prompt automatically detects and strips sensitive data from text pasted into LLM chat interfaces — before it reaches the model. What it strips: • Unix, macOS, and Windows file paths → [LOCAL_PATH] • Custom keywords (project names, codenames, internal terms) → [REDACTED] • Future: API keys, emails, IP addresses Features: • 100% local — zero data leaves your browser • Paste-level interception — no DOM hacks, works across LLM redesigns • Universal mode or per-site selection • Custom matchlist for project names, codenames, or internal jargon • Lightweight, no analytics, no telemetry Privacy: Only requests clipboardRead and storage permissions. No servers, no accounts, no tracking.