AgentsView — See What AI Agents Are Actually Doing

GitHub repo kenn-io/agentsview lets you visualize agent execution traces — not just the final output, but the full path the agent took to get there. You can watch it plan, reason, call tools, backtrack, and sometimes just plain fail in the most predictable ways. It's built by Kenneth with community help, and it's already picking up stars as more teams start shipping agent-heavy apps.

If you've ever wondered why your AI agent repeated the same thing twice or took a weird detour, this is the kind of tool that saves you from guessing. It turns the black box into something you can actually look at — traces, tool calls, reasoning steps, all laid out. It's useful for debugging, for optimizing, and for the quiet realization that your agents are doing exactly what you told them to do, which is sometimes worse than you thought.

Why this matters for us: the more we use AI agents in our apps and workflows, the more we need to see under the hood — not just trust whatever comes out the other side.