Roadmap

What is coming next, in no order and with no dates. An item is here once it is decided; it gets its guide when it ships, marked with the release it came in.

Planned

• Teams: share a project’s docs and plans with people you choose. A team never gives access to a machine.
• Docs and plans on all your nodes: pushed docs coming back down to your other nodes and a team’s, with a conflict kept for a person to resolve, never overwritten.
• Single sign-on (OpenID Connect) for an organisation’s hub.
• Signed releases.
• Agents’ prompts on every AI CLI: reading the question a waiting agent asks on Codex, Cursor and Antigravity, as on Claude Code, and wdx answering them.
• Agents resumed by themselves: an agent running when the machine went down comes back on its own, in the same conversation.
• A plan’s autonomous mode: a plan that settles what stops it (a waiting agent, a conflict, an agent that ended) without waiting for a person.

Shipped

• The hub, in 0.5.0: your machines as nodes of one account, with who is online; remote access with an authenticator code on each machine; --node for commands, attach and TIDE; a project’s docs pushed to the hub; all self-hosted, on Linux or Fly.io, for yourself or an organisation. See Hub and Running a hub.
• Usage and limits, in 0.5.0: each AI CLI’s usage in wdx usage and TIDE’s stats tab, and an agent marked limited while a limit its model draws on is reached. See Agents.

Not planned

Anything not above. Ideas are welcome as issues; Contributing says how an idea becomes a feature.