Raindrop | Monitoring for AI agents

Hey PH! This is Alexis, co-founder of Raindrop.

Your agent fails at 1 am, the traces are in some SaaS dashboard, the harness is on your machine, the eval suite is in a third place, and the cloud doesn’t see any of it.

We are stuck in this cycle. So we made a way out of it. Workshop is the first sensible way to debug your agent locally.

It has two parts: a local UI and an MCP.

Each Span stream is live on a local browser UI and you can replay any agent run with edited signals, models and tools.

MCP lets you create self-healing eval loops. Cloud Code reads your traces, evaluates, and fixes what’s broken.

It’s free, open source, and works with all the agent SDKs you already use.

A command to install: curl -fsSL https://raindrop.sh/install | Bash

Excited to hear what you think!



<a href

Leave a Comment