Orchestrate multiple AI coding agents with your team
Next-gen agent orchestration for AI coding. Multiplayer workspace for Claude Code, Codex, and Gemini.
Agor's session forking and spawning creates conversation genealogy trees where users can branch from any point to try different approaches, keeping all exploration paths visible and resumable. Each fork maintains the conversation history up to that point while allowing independent divergence.
Agor's zone system uses templated prompts that auto-execute when branches are dragged into workflow zones (e.g., 'Ready for Review', 'Needs Tests'). Templates dynamically inject context from git branches, issues, and PRs, eliminating manual copy-paste of URLs and metadata.
Agor uses a 2D spatial canvas instead of linear terminal scrollback, leveraging spatial memory to help teams remember where work is located. Branches and sessions can be arranged by priority, status, or relationships, making project state instantly scannable.
Agor exposes an MCP surface that allows coding agents (Claude Code, Codex, Gemini) to drive themselves through the platform, manage branches, and trigger workflows autonomously rather than requiring manual user intervention.
Agor provides a spatial canvas where teams can fork and spawn AI sessions on isolated git branches, allowing parallel work streams with shared context while maintaining conversation history trees. Real-time multiplayer cursors and comments enable teams to see who is working on what and collaborate asynchronously.