Anthropic

AI safety company and creator of Claude. Publishes engineering articles on agentic infrastructure, model training, and AI systems.

Relevant work in this wiki

• Managed Agents Architecture — Anthropic's approach to scalable agent infrastructure (Brain/Hands/Session decoupling)
• Harness Design Pattern — harness-outside-container pattern
• Durable Session Storage — external append-only session log

Related entities

• Claude — Anthropic's LLM, the "Brain" in Managed Agents
• OpenClaw — third-party agent framework built on top of Claude (compare architecture)