Batty
Rust CLI for running supervised multi-agent coding teams in tmux with YAML-defined roles, isolated git worktrees, and test-gated completion.
Do not bounce yet
Read the fit check, compare one alternative, then decide whether the vendor page is still your best next click.
Quick Verdict
Fast fit check before you leave the page
Make the fit call first. Vendor pages are good at selling, but they rarely tell you where the product is a bad match.
- Terminal-heavy developers already using Claude Code, Codex, or similar coding agents
- Builders experimenting with multi-agent implementation workflows on one repository
- Teams that want stronger supervision and review discipline before parallel agent work hits mainline code
- Adoption is still tiny, so long-term community depth and battle-tested patterns are not proven yet.
- You need to be comfortable with tmux, CLI tooling, and supervising agent workflows instead of expecting a polished hosted UI.
- The hierarchical team model can be overkill for simple solo tasks where one coding agent is enough.
Compare Next
Take one more internal step before the vendor pitch
This is where visitors usually jump out too early. Read one deeper take or open one alternative so the next click is informed instead of impulsive.
Alternative profile
Agent of Empires
Open-source terminal and web session manager for running multiple AI coding agents across isolated git worktrees and sandboxes.
Alternative profile
Claude Squad
Open-source terminal app for managing multiple AI coding agents like Claude Code, Codex, OpenCode, and Aider across isolated workspaces.
Alternative profile
Dirac
Open-source coding agent for VS Code and the terminal with hash-anchored edits, AST-aware refactors, and lean context curation.
Batty is not trying to be yet another shallow wrapper around one coding model. It is a Rust CLI built for people who already know that parallel agent work gets messy fast: branches collide, tasks blur together, and agents claim victory before the codebase is actually safe. Batty answers that by formalizing supervision. You define a hierarchy of roles in YAML, let the tool run them in tmux, keep work isolated in separate git worktrees, and treat passing tests as part of completion rather than an optional afterthought.
Batty is a terminal-native supervision layer for people who want more than a pile of parallel coding-agent sessions. You define hierarchical roles like architect, manager, and engineers in YAML, then Batty runs them in tmux, routes messages between roles, isolates engineer work inside separate git worktrees, and refuses to treat work as done until tests pass. That makes it materially more relevant to serious vibe-coding and agent-coding workflows than the usual launch-week wrapper: it is opinionated about coordination, review discipline, and local inspectability rather than just spawning extra chats.
Why Choose Batty?
Choose Batty if your bottleneck is no longer prompt quality alone but the operational chaos of coordinating several coding agents at once.
Its tmux-native, file-based runtime is attractive for builders who want a local, inspectable control layer instead of a hosted browser abstraction.
Batty is more opinionated about worktree isolation and test gates than many multi-agent experiments, which makes it unusually relevant to serious repo workflows.
The public docs, demo, and release cadence give it more substance than the average launch-week agent orchestrator.
Key Features
Define hierarchical agent teams in YAML, including architect, manager, and engineer roles.
Run each agent inside its own tmux pane while Batty routes role-to-role messages in the background.
Give engineers isolated git worktrees so parallel tasks do not trample the same branch state.
Gate completion on tests instead of trusting agents to self-certify that work is done.
Drive work from Markdown kanban boards and emit file-based runtime state you can inspect or diff.
Stay agent-agnostic at the execution layer, with public positioning around Claude Code, Codex, and Kiro.
Pros & Cons
- Batty tackles the real multi-agent coding problem: supervision, isolation, and merge discipline, not just spawning more sessions.
- The local-first file-based runtime makes it more auditable than hosted black-box coordination layers.
- YAML team definitions and tmux-native execution give experienced terminal users a workflow they can actually adapt.
- Official docs, a public demo, and active releases make it more credible than most ultra-early agent dashboards.
- Adoption is still tiny, so long-term community depth and battle-tested patterns are not proven yet.
- You need to be comfortable with tmux, CLI tooling, and supervising agent workflows instead of expecting a polished hosted UI.
- The hierarchical team model can be overkill for simple solo tasks where one coding agent is enough.
- Batty orchestrates other agents; it does not remove the need to manage credentials, model access, or human review.
Detailed Use Cases for Batty
Use Batty when several agents need to work on one repository but you want explicit role boundaries and operator oversight instead of unmanaged session sprawl.
Batty is well suited to teams that want each engineer agent on its own isolated git worktree so parallel changes do not fight over the same branch state.
If agents keep declaring work complete while leaving the repo broken, Batty offers a more disciplined pattern by making tests part of the completion path.
Batty is a fit for builders who want their agent orchestration layer to stay in tmux, YAML, and versionable local files instead of disappearing into a hosted product.
Who Should Use Batty?
Terminal-heavy developers already using Claude Code, Codex, or similar coding agents
Builders experimenting with multi-agent implementation workflows on one repository
Teams that want stronger supervision and review discipline before parallel agent work hits mainline code
Local-first users who prefer inspectable files and git-native state over browser-only coordination tools
Perfect For
Running several coding agents on the same repository without shared-branch collisions.
Experimenting with architect-manager-engineer team structures for larger implementation tasks.
Adding test-gated discipline to vibe-coding workflows that were previously too chaotic to review cleanly.
Keeping agent runtime state local and inspectable instead of moving supervision into a hosted browser-only layer.
Technical Details
Batty Comparisons & Alternatives
Batty review
Batty vs Claude Squad
Batty vs Superset
tmux multi agent coding tool
git worktree AI coding orchestrator
test gated coding agent workflow
Developers compare Batty with other vibe coding tools when they need a better workflow fit, not just a better landing page.
Direct Competitors
Claude Squad
Superset
Agent Of Empires
Dirac
Similar Tools You Might Like
Official open-source MCP server and CLI that lets coding agents control, inspect, and debug a live Chrome browser with real DevTools context.
Open-source terminal coding agent with a multi-agent architecture, custom agent workflows, and SDK support for repo-aware code generation.
Open-source spec-driven framework that adds lightweight proposal, design, task, and archive workflows to modern coding agents.
Alternative Tools to Consider
Open-source terminal and web session manager for running multiple AI coding agents across isolated git worktrees and sandboxes.
Open-source terminal app for managing multiple AI coding agents like Claude Code, Codex, OpenCode, and Aider across isolated workspaces.
Open-source coding agent for VS Code and the terminal with hash-anchored edits, AST-aware refactors, and lean context curation.
Minimal open-source terminal coding agent focused on extensibility, tree-structured sessions, and shell-native repo workflows.
Worktree-based macOS editor for running and reviewing multiple CLI coding agents like Claude Code, Codex, Gemini CLI, and OpenCode in parallel.
Do one more comparison before you commit to Batty
Strong picks usually survive one more internal check. Read deeper, compare a neighbor, then leave for the vendor page if the fit still holds.