Hi, I'm Milan, a solo developer and computer engineering student from India. I'm building AgentPulse, a platform for managing long-running AI agents with runtime controls, observability, and workspace management. Interested in AI infrastructure and developer tools.
This is an interesting approach. I like the idea of describing a strategy in natural language instead of wiring everything together manually. How do you handle ambiguity when a prompt could be interpreted in multiple ways?
Nice idea. I'm curious how you detect that two meetings from different sources are actually the same meeting. Is it mostly based on timestamps, titles, or some fuzzy matching heuristic?
Nice idea. One thing I'm always concerned about with secrets management is recovery and auditability. How are you handling encryption and access control?
Congrats on the progress. Reaching users across so many countries so quickly is impressive. What acquisition channel ended up working best for you in the early days?