Ex-Moonshot Dev Launches Raft 1.0 for Multi-Agent Collaboration

Raft 1.0, developed over 4 months by former Moonshot engineer RC, has been officially released. The product is designed as a "team mode" shared workspace for multiple AI agents, aiming to solve the pain point of losing context when users manually switch between terminals, sessions, and skills in current workflows. This marks a shift in the AI bottleneck from prompt engineering to multi-agent management.

Confirmed

• Background: Creator RC previously worked at Moonshot on Kimi CLI and built a chatbot in 2015. Raft 1.0 has undergone a 4-month public development period.
• Core Features: Raft 1.0 allows users to place multiple specialized agents into a unified workspace, coordinating them through a Slack-like group chat interface.
• Technical Mechanism: The product introduces a shared memory mechanism, enabling different agents to collaborate on tasks with clear divisions of labor without requiring users to constantly switch contexts.

Why it matters

Observers like @heyabusiddik note that the future of AI likely lies not in a single "super assistant," but in multiple specialized agents with clear divisions of labor, shared memory, and collaborative workflows. Raft 1.0 aligns with this trend, transforming the multi-agent collaboration experience from tedious terminal switching into an intuitive team communication model, directly addressing the management pain points of current agent workflows.

2026-07-23 ~ 2026-07-24 · 7 related posts

• Raft 1.0 launches a shared workspace for AI agents, built by Moonshot Kimi CLI creator RC — CodeByPoonam · 2026-07-23
• Raft 1.0 launches team mode for multi-agent workflows — hey_abusiddik · 2026-07-24
• Raft 1.0 launches team mode for agents with shared memory and role-based workspaces — hey_abusiddik · 2026-07-24
• Raft 1.0 launches a team-mode workspace for running multiple agents — hey_abusiddik · 2026-07-24

3 near-duplicate retellings: hey_abusiddik · hey_abusiddik · hey_abusiddik