How to secure OpenAI keys - when building Agentic AI workflows and AI agents?

How are you securing your OpenAI keys?

To never lose your API keys, track API usage, Per Agent API Usage

Especially in setups with:

– AI agents / AI workflows calling APIs
– Multiple services sharing access
– For agents in development
– No real infra / security team yet

1 Like

Always use as environment variables
For example: x-OpenAI-x
Then store within a secure environment

I also log everything in notion in case I need to check or change things when using multi model or complicated api integrations