Raft introduces a new platform designed to seamlessly integrate persistent AI agents into team workflows, enabling continuous, context-aware collaboration alongside humans while maintaining control over data and governance.
Raft is positioning itself as a different kind of collaboration platform: one built not just for people, but for people and persistent AI agents working together as if they were members of the same project team. According to the company’s website, the system gives agents memory, defined roles and continuing responsibilities, while keeping the human user in control of direction and final decisions.
The practical aim is to reduce the reset that often comes with using AI as a series of isolated prompts. Raft’s documentation says agents can join shared channels and threads, review prior context, claim tasks and return to work later with an understanding of earlier discussions. That design is meant to make them more useful for ongoing work such as code review, drafting, triage and feature planning, where context tends to matter as much as raw output.
A notable part of the platform is its runtime flexibility. Raft says it can work with AI systems that users already subscribe to, including Claude Code and Codex CLI, and can also connect external agents running on a user’s own machines. In effect, the platform acts as the coordination layer, stitching communication, reminders and task ownership into one workspace while leaving the underlying model or environment choice to the user.
Raft also emphasises control over data and governance. The company says local workspace content remains under user control, with only necessary metadata, tasks and messages shared across the system. Its published pricing suggests a freemium model aimed at smaller teams, with a free tier that includes channels, tasks, agent reminders and limited history, a Pro plan pitched at about $8.80 per human seat a month when billed annually, and an Enterprise offering with private deployment, single sign-on and granular access controls. That mix makes the product most relevant to teams looking to formalise human-AI collaboration without locking themselves into a single model provider, though its own materials indicate that organisations with strict compliance needs may want to weigh the platform’s current limitations carefully.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
