Building a Scalable Email Management Agent: Best Practices Needed
Hey Lindy Community 👋 Quick caveat up front: I’m very much a beginner here, so please be gentle 😅 I’ve been using Lindy to build a small agent that solves a personal problem for me. In simple terms, it acts like an inbox manager:
Reads my emails
Downloads attachments
Figures out what’s important
Adds events to my calendar when relevant
Sends me a daily TL;DR summary
I showed it to a few people and there’s some interest in using it themselves . This is where I’ve hit a wall. My very naive first thought was to just clone the agent and plug in another user’s email address. But obviously that means needing Gmail/Outlook permissions for their inbox, which opens up a whole set of challenges (privacy, auth, trust, etc.). Ignoring privacy for a moment (I know), my main questions are:
How do you actually handle Gmail/Outlook OAuth approvals for other users in a scalable way?
Is there a common pattern for this when turning a personal agent into something others can use?
One alternative idea I had was:
A “stateless” or “brainless” agent that just reads emails / downloads attachments and passes the raw output on
A second, user-specific agent that analyses the data, stores memory, sends summaries, and creates calendar events
But I’m not sure if that’s sensible or over-engineering things. Where I’m currently stuck:
Data privacy & isolation between users
Avoiding data leakage across agents
Gmail / Outlook authentication & consent flows
How to automate all of this so it actually scales with new users
If anyone has been down this road before, or can point me to best practices / "don’t do this” advice, I’d massively appreciate it 🙏 Thanks in advance!
