I Rebuilt Hermes in Claude Code (It’s Ridiculously Good)

Uploaded: 2026-05-23
The video discusses the rapid adoption of the Hermes AI system on GitHub, highlighting its innovative features and the potential pitfalls of using off-the-shelf software like Open Claw. The creator shares their experience of rebuilding elements of Hermes into a custom setup that offers better understanding, modularity, and scalability, ultimately emphasizing the long-term benefits of a tailored approach.

channel	Simon Scrapes
top	Fan_month

Add to wishlistAdded to wishlistRemoved from wishlist 0

Add your review

SKU: wdc1OFWDxlU

Description
Statistics
Additional information

Explores rebuilding Hermes features inside Open Claw, comparing off-the-shelf agent trade-offs with a custom modular agentic OS. Focus on assumptions, security, memory, identity and maintainability for multi-client workflows.

– Hidden costs: inherited assumptions, security gaps, and audit losses.
– Memory & identity: injected context plus semantic recall for multi-brand use.
– Skill systems: modular single-source components to avoid duplicate skills.

Quotes:

Hermes went from zero to 40,000 GitHub stars in 46 days.

The self-learning loop has no external guardrails — it writes skills, then grades its own homework.

Faster to start with Hermes, but faster to scale with your own setup.