Jackalope unifies parallel AI coding agents
Jackalope is a cross-platform desktop workspace for running Codex, Claude Code, Grok, OpenCode, Kimi Code, and Antigravity tasks side by side in isolated Git worktrees. It centralizes project context, account profiles, MCP tools, recurring tasks, and code review while remaining in early access.
Jackalope targets the coordination gap emerging around AI coding agents: the agents already exist, but managing their parallel output remains messy. Its strongest idea is treating agent work as reviewable project work rather than disposable terminal sessions.
- –Isolating tasks in Git worktrees enables parallel exploration without immediately destabilizing the main branch
- –Shared project guidance, dependencies, messages, and review history could reduce duplicated context across agents
- –Capacity-aware routing and separate work or personal profiles address practical subscription and quota friction
- –MCP, browser automation, computer use, and recurring tasks extend the workspace beyond code generation
- –Jackalope does not provide model access; developers still bring their own CLIs, accounts, permissions, and provider costs
DISCOVERED
1h ago
2026-09-11
PUBLISHED
7h ago
2026-09-11
RELEVANCE
AUTHOR
Caden Sumner
