Submit a tool Sign in
Live

Redaxa

AI prompt privacy scanner and redaction tool that detects credentials, personal data, and secrets before they reach large language models.

Redaxa is a prompt privacy scanner and data redaction tool designed to protect confidential information before it reaches AI models. As developers and teams rely on assistants like ChatGPT, Claude, Gemini, and Perplexity for coding and writing, private records frequently leak through copy-pasted text. Redaxa acts as a security filter, catching credentials, financial data, and personal details before transmission.

The scanner flags exposed API keys, database connection strings, passwords, private keys, emails, and bank IBANs. When sensitive elements are found, Redaxa highlights them and produces a sanitized version using neutral placeholders. This preserves the semantic context of your query, allowing AI models to generate helpful answers without seeing actual private values.

Redaxa is available through a web app, a Windows desktop client, and a Chrome extension that runs inside popular AI chat interfaces. Developers can also integrate the dedicated Scan API into CI/CD pipelines, terminal scripts, or internal AI middleware. Business plans add shared organization terms and custom enforcement rules to warn, redact, or block sensitive submissions.

The platform also features a public repository scanner that checks GitHub projects for leaked tokens and reviews dependencies against known vulnerability databases. Built with a zero-retention design, server-side prompt checks are evaluated over encrypted connections and discarded immediately without logging.

Features

  • Prompt checks for potentially sensitive information
  • Suggested redactions to inspect before sharing
  • Documented repository-review workflows in the Windows app
  • Audits public repositories for leaked secrets and cross-references package lockfiles against the OSV open-source vulnerability database
  • Replaces exposed credentials and private data with structured category placeholders so AI tools can still answer accurately

Use cases

  • Review a client-related prompt before sharing it with an AI tool
  • Inspect possible credential exposures in a supported repository workflow
  • Review a redacted draft and decide what information to share
  • Team Privacy Policy Enforcement
  • DevOps teams integrate the Scan API into build scripts to prevent accidental commits of customer PII and credentials to external AI services

Screenshots

Redaxa screenshot Redaxa screenshot

Frequently asked questions

Does a clean result guarantee that my information is safe?

No. Detection can miss sensitive information or produce false positives. Always review the original text and the suggested redactions yourself.

Can Redaxa scan code repositories for exposed credentials?

The Windows application provides documented repository-review workflows for possible exposures. Review its supported scope and permissions on the official website; results are not a complete security assessment.

Is there a free way to test Redaxa before subscribing?

The official website offers free entry options separately from optional paid plans. Check the current allowance, eligibility, processing disclosures and purchase terms there.

Who operates the Redaxa service and where is account data stored?

Redaxa is developed and operated by software engineer Aurelio Avila, with account authentication and billing metadata managed through Supabase in the European Union and Stripe.

What are the request limits for the Redaxa Scan API?

The Scan API accepts prompts up to 20,000 characters in length and permits rate limits of up to 120 checks per minute per account, returning structured JSON findings with policy action decisions.

What we verified

Public API
Yes (Offers a REST Scan API authenticated via Bearer tokens for programmatic text redaction, custom category filtering, and metadata-only activity querying)
Self-hostable
No (Cloud-hosted architecture deployed on Vercel and Supabase EU, with a native Windows desktop application and Chrome browser extension)
Free tier
Yes (Free tier includes 5 prompt checks per day on web and Windows, unlimited on-device checks in the Chrome extension, and 3 public GitHub repository checks per day)
Pricing model
Subscription (Monthly and annual recurring billing with a 7-day free trial on Pro and Business tiers)
Price
Free tier is €0; Pro plan is €7.99 per month or €79.90 per year; Business plan is €14.99 per user per month or €149.90 per user per year (supports 1 to 3 users)
Data region
European Union (Supabase database and authentication hosted in the EU region; application hosted on Vercel; prompt text is processed ephemerally in memory and discarded without storage)
Integrations
Chrome Web Store extension (integrates into ChatGPT, Claude, Gemini, and Perplexity), Windows desktop application, GitHub (public repository checking), OSV.dev (vulnerability lookup), Supabase (EU), Stripe, Resend

Could not be verified

Support for private GitHub repositories (explicitly documented as unsupported on the web checker; local git analysis is limited to Windows app with Pro or Business)
Underlying regex engines or specialized natural language processing libraries used in the server-side credential detection pipeline

0 comments

Sign in to leave a comment.

No comments yet. Say something useful and be the first.