Show HN: Xyzzy – AI teamwork in one Python process, with a tamper-evident log (github.com)

XYZZY, a new tool announced by a development team, enables real-time AI collaboration within a single Python process, equipped with a tamper-evident event log. Designed for use by teams rather than individual users, XYZZY allows multiple humans and AI agents to share a digital workspace where they can interact, review outputs, and make decisions based on recorded evidence. The platform supports multi-agent orchestration, meaning agents can operate in parallel while humans maintain control over the final actions, creating a "human-in-the-loop" environment. This development is significant for the AI/ML community as it addresses a critical gap in existing tools, which typically cater to solo work rather than teamwork. By offering features such as versioning of documents, artifact tracking, and proof of actions through a hash-chained event log, XYZZY ensures transparency and traceability in the decision-making process. The tool employs SQLite for data persistence and provides a self-hosted option, giving teams the flexibility to integrate various AI models without depending on external APIs. Overall, XYZZY represents a major step toward collaborative AI environments, enhancing productivity and accountability in team-based AI initiatives.