Problem solved

Developers building complex software need to coordinate multiple AI coding agents to tackle different aspects of a problem simultaneously, but manual orchestration leads to context bleeding, workspace conflicts, and fragile inter-agent communication. tuttisolves this by providing a unified CLI that automates parallel agent execution with isolated git worktrees and a typed artifact pipeline, ensuring each agent operates in a sandboxed environment while seamlessly sharing structured outputs.

Use cases

• Parallel code generation where multiple agents work on different modules or features of a project simultaneously.
• Automated code review and testing by dispatching agents to run linters, tests, and security scans in parallel
• Bootstrapping large codebases by dividing tasks among agents that handle configuration, scaffolding, and documentation.
• Executing similar prompts across multiple git branches to compare agent performance or code variations.
• Integrating into CI/CD pipelines for automated multi-agent workflows that build, test, and deploy in concert.
• Orchestrating agent-based refactoring sprints where each agent is assigned a refactoring task with strict isolation.

Audience

• Developers
• ML Engineers
• Researchers

Pros

• Config-driven workflows ensure reproducibility and easy version-control of agent orchestrations.
• Git worktree isolation eliminates workspace conflicts, enabling true parallel agent execution.
• Typed artifact flow provides structured, type-safe communication between agents.
• Lightweight and fast due to Rust implementation, suitable for high-throughput environments.
• Open-source and extensible through integrations and custom skills.

Cons

• Early-stage project with limited community support and documentation beyond the repository.
• Requires familiarity with configuration files and agent ecosystems, which may present a learning curve.
• Current integrations may be limited to specific AI coding tools, reducing flexibility.