Human control plane
The human defines objectives, assigns responsibility, sets boundaries, observes execution, and accepts results.
Wega Labs infrastructure · October
Infrastructure for supervised AI collaboration.
Wega Labs projects require multiple AI agents to share context, divide work, and expose their activity to a human supervisor. October is the collaboration layer we are building for that work. Its current interface is a spatial desktop canvas.
Current formDesktop spatial canvas for coding agents
Collaboration problem / 01
Multiple agents need shared state, coordination, and human control.
Agent processes run in separate terminals, chats, editors, repositories, and provider applications. They do not share a consistent model of tasks, ownership, dependencies, or current system state.
October provides a common control and coordination layer while keeping the human responsible for scope, review, and final decisions.
System model / 02
A shared collaboration layer across agent runtimes.
Agent coordination
A shared protocol provides peer discovery, messages, task ownership, canvas context, and execution status.
Execution adapters
Agent runtimes work in isolated repositories and expose terminals, diffs, services, and previews to the control plane.
Models, providers, editors, and agent runtimes remain independent. October coordinates them through a common human control plane and agent protocol.
Coordination protocol / 03
Agents communicate through a shared bus.
The october-bus exposes peer discovery, messaging, relevant canvas state, task ownership, and execution status through MCP. It gives agents a shared operational context without granting unrestricted access to one another.
HumanIntent + direction
→OrchestratorResolve names, screens, ownership
→Screen A
Backend
Tests
October bus
discover peers · inspect canvas · message agents · report status · preserve provenance
System status / 04
Current implementation and remaining work.
Built + shipping
Current desktop- Spatial infinite canvas✓
- Route detection + live previews✓
- Local coding-agent harnesses✓
- Integrated terminals✓
- Git worktree isolation✓
- Variant previews + PR flows✓
- Otto voice control✓
- Cross-platform CI gate✓
Agent roster
Local harnesses01Claude Code02Codex03Cursor04Grok05Gemini06opencode07Hermes08Cline09Pi
All listed agents can run locally. Their october-bus integrations are not equivalent.
In development
Named agent control
Direct a specific agent by name or by its assigned screen.
Protocol parity
Claude Code currently has the deepest bus integration. Other agents need the same discovery, messaging, and status functions.
Concurrent editing
Add explicit ownership, presence, provenance, and synchronization for simultaneous work.
Local and remote agents
Expose local and remote agents through the same control and coordination interfaces.
Human supervision / 05
The human defines scope, observes execution, and accepts results.
Visibility
Agents, tasks, terminals, services, and output remain visible in one workspace.
Responsibility
Each task has an assigned agent, bounded scope, dependencies, and reported status.
Isolation
Git worktrees separate concurrent sessions before changes are reviewed and merged.
Provenance
The system records which agent performed work, which files changed, and where output was produced.
Supervised execution cycle / 06
1Define scope
→The human selects repositories, objectives, constraints, and approval boundaries.
2Assign
→Tasks are assigned to named agents with visible ownership and dependencies.
3Coordinate
→Agents exchange relevant context and status while executing in isolated workspaces.
4Review
The human inspects previews, diffs, tests, terminals, and provenance before acceptance.
Relationship to December / 07
December is the research system. October is collaboration infrastructure.
Building December requires coordinated work across simulation, software engineering, experiments, analysis, operations, and review. Independent agents need to exchange context without making responsibility unclear.
Today, October coordinates coding agents working on software. Broader coordination infrastructure for different agent roles and Wega Labs projects is still in development.
Current form factor / 08
The first interface is a spatial canvas.
A persistent canvas lets humans and agents refer to the same screen, task, position, and local context. Other interfaces can use the same coordination layer later.
October by Wega Labs
Coordinate AI work under human supervision.
The current desktop application uses a spatial canvas for coding agents, live software, terminals, tasks, and review.
Visit october.dev ↗