Per-project template variables
Use {pm} for the project's package manager (npm, pnpm, yarn, bun), {script:NAME} for package scripts, {branch} for Git branch, and {name} for project name.
Multi-terminal broadcast
Broadcast Terminal: Run commands across all repositories at once.
Execute shell commands across multiple projects, project groups, and Git worktrees simultaneously. Use per-project template variables, live-sync keystrokes, and fix failures with AI.
The problem
Running the same command in ten terminal tabs is tedious and error-prone.
In multi-repo systems and microservices, daily tasks require running commands across multiple repositories: checking out branches, running migrations, installing dependencies, building packages, or restarting services.
Managing ten separate terminal windows or running shell loops across directories means losing output clarity, missing errors buried in scrollback, and struggling to handle repos with different package managers or scripts.
Flame Broadcast Terminal arranges your projects into a responsive tile grid. Run templated commands across all checkouts in one shot, see every log stream side-by-side, and let AI resolve failures.
Multi-terminal broadcast gridTarget projects, fan out commands, stream output in real time, and fix failures with AI.
How Flame solves it
From targeting repositories to verified execution in five steps.
- 01
Select target projects
Target a saved project group, active worktree fleet, all engaged repositories, or an ad-hoc selection of local checkouts.
- 02
Compose templated commands
Write command strings with dynamic tokens like {pm} (package manager), {script:dev}, {branch}, and {name} that resolve accurately per project.
- 03
Choose execution policy
Run commands in parallel or sequentially, set concurrency limits, and choose whether to halt immediately when a project command exits with an error.
- 04
Live-sync interactive input
Enable live-sync mode to fan out keystrokes and control multiple terminal tiles simultaneously with real-time PTY terminal emulation.
- 05
Investigate failures with AI
When a service, linter, or test fails, send the terminal output directly to an AI agent to diagnose the root cause and generate a fix in place.
Execution power
Built for complex multi-project development.
Interactive live sync
Send keystrokes, Ctrl+C signals, and interactive responses to all member tiles simultaneously over high-speed WebSocket connections.
Dangerous command guards
Built-in danger confirmation popovers intercept destructive commands (like rm -rf, git reset --hard, or DROP TABLE) before they fan out.
Concurrency & error policy
Control execution speed with configurable concurrency limits and choose whether to halt immediately on error or let remaining jobs finish.
Saved broadcast commands
Store frequent multi-project routines with scope annotations (global, project group, or repository-specific) for one-click re-runs.
One-click AI error triage
When a command exits with a non-zero code, hand the full PTY output directly to an AI agent to analyze error traces and generate fixes.
Product walkthrough
Watch command broadcasting across multiple repositories.
Official Flame demo
Broadcast Commands Across Repositories & Fix Failures with AI
Run terminal commands across every project at once and let AI investigate and fix failures in place.
Connected workspace
Integrate broadcast terminal with your service stack.
Broadcast Terminal pairs directly with multi-repository workspaces and microservices development. You can define group run configurations that launch backend services, database migrations, and frontend dev servers together.
When working with parallel branches or reviewing PRs, broadcast commands across active Git worktrees to run test suites or compile bundles in isolated checkouts.
FAQ
Broadcast terminal questions
Stop running the same terminal commands in a loop.
Download Flame to broadcast commands, live-sync interactive keystrokes, and coordinate all your repositories.
Free to use on macOS, Windows, and Linux.