Try it out! (Just double-press ⇧ Shift after typing your prompt to activate.)
I found myself constantly going back to manually redact sensitive inputs from my prompts, and quite frankly, I'm too lazy to keep doing that by hand.
- LLM Enabled: Intelligent PII detection and redaction using local LLMs.
- In-Browser: All processing happens locally within your browser—no data leaves your device.
- Near-Instant Redaction: Optimized for speed to minimize friction.
- Response Rehydration: Automatically restores redacted information in Gemini's responses for a seamless experience.
- Framework: React, TypeScript, WXT
- UI: Mantine
- AI: WebLLM (running Llama-3-8B-Instruct in browser)