We're building the first secrets manager designed specifically for AI-assisted development.
THE PROBLEM:
You're deploying your app and need help from Claude or ChatGPT.
The AI asks: "I need your database credentials, API keys..."
What do you do? Most developers paste their secrets directly into the chat.
This is a massive security risk. Those secrets are now in:
• Chat logs (potentially stored indefinitely)
• Training data (if not properly filtered)
• Your conversation history (accessible to anyone with your account)
But what's the alternative? Work without AI help on deployment tasks?
THE SOLUTION:
We're building AgentSecrets - a CLI tool that lets AI assistants help with deployments WITHOUT ever seeing your actual secrets.
How? Zero-knowledge architecture. AI manages the workflow, not the values.
What secrets management challenges do you face? Drop a comment.
Top comments (0)