Show HN: Add human approval layer to monitor and guide LLM agents in production

Hi HN!

In the past month, I have been building Phantasm. Phantasm offers toolkits to create a human approval layer for AI agents which allows teams to monitor and guide the workflow of their AI agent in real-time. It is fully open-source and comes with a server and a custom web-based dashboard.

I built this because I was working with a team building AI-powered CRM with a feature to update a deal data based on a sales call automatically using LLM. We received feedback that the result was not satisfying which lead us spending weeks to build similar system internally.

Phantasm is still very early and I would love to hear feedback from the community to improve the project!

Thank you in advance.