Show HN: Sayou: Open-source agent workspace with versioned files and MCP tools (github.com)

Sayou, an open-source workspace for AI agents, has emerged as a game-changing tool designed to replace conventional databases with a file-system-oriented context store. It allows AI agents to retain knowledge across sessions, enabling persistent reasoning and context accumulation. Key features include versioned files that are auditable and structured with YAML frontmatter, making it possible to track changes over time through an append-only version history. This approach significantly enhances the capabilities of AI agents by allowing them to manage and recall information seamlessly across multiple interactions. The significance for the AI/ML community lies in Sayou's ability to bridge the gap between transient chat interactions and long-term memory. By leveraging a file-first, SQL-compatible system, Sayou integrates smoothly with various agent architectures, providing tools like version control, file searching, and knowledge management. Its compatibility with existing AI frameworks, such as a REST API and Python libraries, allows easy implementation across projects. Consequently, Sayou not only empowers AI systems to remember information effectively but also introduces a new way to evaluate and benchmark agent performance, potentially transforming workflows in professional settings where knowledge continuity is paramount.