PromptVigil

About

PromptVigil helps you use AI tools more safely.

  It protects supported AI sites like ChatGPT, Claude, and Copilot by scanning prompts, responses, links, and supported file uploads before risky content causes harm.

  What PromptVigil does:

  - Warns before sensitive data is shared with AI tools
  - Detects exposed secrets such as API keys and tokens
  - Flags suspicious links and scam-style content
  - Detects common prompt injection and unsafe instruction patterns
  - Redacts risky content when a safer retry is possible
  - Shows an activity log with redacted event details and action history
  - Provides local-first protection controls and policy settings

  How it works:

  - PromptVigil analyzes supported AI interactions locally in the browser by default
  - Raw prompts and raw AI responses are not sent by default
  - Raw text logging is off by default
  - Anonymous telemetry is optional and does not include raw prompt or response text

  Who it is for:

  - Individuals who want safer daily AI use
  - Teams that want guardrails for AI prompts and file uploads
  - Organizations that want visible, user-facing AI interaction protection

  Current protection areas:

  - Sensitive data and payment-like values
  - Secrets and credentials
  - Suspicious links
  - Prompt injection patterns
  - Fraud and scam-style content

  PromptVigil is designed to reduce accidental oversharing and risky AI interactions without requiring users to change how they work.