Prompt Seal – Hide API Keys & Secrets from ChatGPT, Claude, Gemini
Overview
Stops API keys, tokens and passwords from reaching AI chats. Redacts only the secret, keeps your config readable. Runs locally.
🦭 Prompt Seal — Stop leaking API keys and secrets to AI chats Pasting a .env file, a Kubernetes manifest or a stack trace into an AI chat is fast — and one of the easiest ways to leak a production credential. Prompt Seal is a free, local safety net for developers: it catches API keys, tokens and passwords before you hit Send. 🔒 WHAT IT DOES • Warns while you type — a small badge above the chat box shows how many possible secrets are in your draft • Pauses the send — on Enter or Send, Prompt Seal checks your message locally and shows what it found (values stay hidden) • You decide — Redact & Send, Cancel, or Send Anyway ✂️ HIDES THE SECRET, KEEPS THE CONTEXT Only the sensitive value is replaced. Variable names, hosts and structure stay, so the AI can still help you debug: DATABASE_URL=postgres://admin:[REDACTED]@db.internal:5432/app AWS_SECRET_ACCESS_KEY=[REDACTED] ⚡ WHAT IT CATCHES • Cloud and AI provider API keys • Source-control, CI/CD and package-registry tokens • Payment, messaging and monitoring service keys • Passwords in .env, JSON, YAML and TOML files and in command-line flags — even short ones • Passwords inside database connection strings • Kubernetes Secrets, private keys, JWT and Bearer tokens • Card numbers and IBANs, validated by checksum to reduce false alarms • Unknown high-entropy secrets next to names like "password", "token" or "secret" ✅ FEWER FALSE ALARMS Public SSH keys, certificates, plain e-mail addresses, hashes and placeholders such as ${DB_PASSWORD} are left alone. 🛡️ BUILT TO FAIL SAFE • Repeated Enter or clicks can't slip through while you review • Other buttons in the chat never trigger a send • If a check can't finish, sending is paused instead of passing silently • After redaction, the text is read back and checked again before sending 🎨 FEELS NATIVE • The warning and badge follow the site's light or dark theme • Each finding shows its line number, never the secret itself • Optional auto-redact, custom badge colour and popup themes • 12 interface languages, detected automatically 🔐 PRIVATE BY DESIGN • Scanning runs locally inside the extension — your prompts are never uploaded, stored or logged by Prompt Seal • No account, no servers, no analytics, no telemetry • Minimal permissions: only "storage", plus access to the supported AI chat sites 💬 WORKS ON ChatGPT, Claude, Gemini, Microsoft Copilot and Perplexity. ⚠️ HONEST LIMITS Prompt Seal is a best-effort safety net, not a guarantee. It can miss secrets or flag harmless text, and the AI website can read what you type before you press Send. Files, images and voice are not scanned. If a real credential was sent, revoke and rotate it. 🌐 Website & FAQ: https://promptseal.vercel.app 🔏 Privacy: https://promptseal.vercel.app/privacy ✉️ Support: contact@alekseialeinikov.com
0 out of 5No ratings
Details
- Version1.2.0
- UpdatedNovember 17, 2026
- Size217KiB
- LanguagesEnglish
- DeveloperWebsite
Email
contact@alekseialeinikov.com - Non-traderThis developer has not identified itself as a trader. For consumers in the European Union, please note that consumer rights do not apply to contracts between you and this developer.
Privacy
This developer declares that your data is
- Not being sold to third parties, outside of the approved use cases
- Not being used or transferred for purposes that are unrelated to the item's core functionality
- Not being used or transferred to determine creditworthiness or for lending purposes
Support
For help with questions, suggestions, or problems, please open this page on your desktop browser