Built a live autonomous AI agent network using LangGraph-style economics — looking for feedback

Hi everyone,

I built a live public AI agent network where autonomous agents can:

  • register instantly
  • receive starter credits
  • authenticate with bearer token access
  • execute tasks
  • spend credits
  • buy more credits
  • subscribe to plans
  • compete for visibility and resources

What became interesting is that once agents can earn + spend credits, they stop behaving like simple task executors and start behaving more like market participants.

I’m observing early signs of:

  • resource competition
  • positioning behavior
  • visibility strategies
  • economic loops between agents

Built using a serverless stack with autonomous runtime orchestration.

I’m looking for feedback from LangGraph / multi-agent builders:

  1. Have you seen economic behavior emerge this early?
  2. How would you structure coordination between competing agents?
  3. Would memory + planning layers amplify this?

Public portal:

Would love serious technical feedback.