skip to content
~/docs~/roadmap~/integrations~/skills~/faq

roadmap

What is shipped, what is next.

Aethereum keeps your team's AI coding agents in sync across every machine: contracts, intent, decisions, and collision alerts. Here is what it does today, what we are building now, and what we are exploring. Everything shipped is free, no signup, one command.

  1. shipped · 8
  2. in progress · 3
  3. exploring · 3

shipped

  1. Cross-machine shared brain

    Every agent on every teammate's machine reads and writes one team context: interface contracts, intent, decisions, and a project brief. Metadata only, never your source code.

  2. Interface-contract negotiation

    An agent proposes a new contract shape, dependents accept or push back with a reason, and it is finalised, all before code is written. A breaking change can park for a human's approval.

  3. Enforced contract gate at the git boundary

    `aethereum check` runs as a git pre-commit hook and flags a commit that touches an interface a teammate is mid-negotiation on, or one marked frozen. Advisory by default, or set it to block. Fully fail-soft: offline or unconfigured never blocks a commit.

  4. Targeted collision alerts

    Register the interfaces you consume with depend_on, and you are alerted only when those change, not about every contract in the room.

  5. Live coordination for native agents

    Claude Code gets automatic context and collision alerts through hooks. Any MCP agent (Cursor, Codex, Windsurf, Cline, Zed) blocks on await_team_events to hear about a teammate's change live.

  6. Mission control + a living ruleset

    Set one standing directive the whole fleet obeys, a shared backlog of tickets, and a versioned team ruleset that propagates to every agent.

  7. Integrations

    GitHub PR comments, Slack and Discord alerts, two-way Linear ticket sync, outbound signed webhooks, an agent-to-agent (A2A) endpoint, and a public read API are all live.

  8. Obsidian-grade dashboard

    A live room graph, presence and intent, mission control, an audit log, and per-panel error isolation. Free during the beta.

in progress

  1. Prevented-collision analytics

    We record every collision averted before code shipped. Surfacing that as the headline metric on the dashboard and in the demo, so the value is measured, not just claimed.

  2. Dynamic AGENTS.md propagation

    The team ruleset ships today; a one-command `sync-rules` that rewrites each machine's AGENTS.md from the room's current rules is next.

  3. Hosted read-path performance

    Folding the team-context read into a single server-side query so large, long-lived rooms stay fast.

exploring

  1. End-to-end encryption

    Today we never store your source code, only the contracts and decisions you publish. E2EE for an enterprise tier is on the roadmap; it is not available yet, and we will say so plainly until it is.

  2. Fork and merge reconciliation

    When two machines genuinely fork the same contract, a merge flow that pairs both shapes for a human to reconcile.

  3. Beyond MCP agents

    A paste-back loop so browser builders (v0, Lovable, Bolt) and other MCP-less agents can join a room too.

full changelog → every tool →

Free, no signup, one command.

get started →