ProtocolProduction ReadyMCP

MassGen

Name: MassGen
Rating: 3.1 (1039 reviews)
Author: massgen

by massgen

CLI-driven multi-agent orchestration with scaling and failure handling

Python

Updated May 28, 2026

1.0k

Stars

160

Forks

View on GitHub

Overview

Orchestrates large-scale multi-agent workflows from the terminal, letting autonomous agents and frontier models collaborate and iterate Orchestrator-Worker Pattern. Uses a CLI-first, test-time-scaling approach to spawn, coordinate, and retry agents with tool-calling and conversational handoffs. Distinctive features include terminal UI controls, parallel agent execution, and hooks for model/tool selection and failure recovery. It also supports cross-agent strategies via Agent-to-Agent Protocol (A2A).

Key Benefits

As agents scale from single-model assistants to networks of specialists, you need systems that manage delegation, retries, and emergent failure modes. MassGen provides the operational layer for running many agents and observing how they interact, which is crucial for building an agent track record and surfacing trust signals, guided by patterns like the Agent Registry Pattern. Until teams adopt repeatable multi-agent orchestration patterns, evaluating agent-to-agent behavior and reliability will remain ad hoc.

Target Use Cases

Teams and researchers who need to run, iterate, and observe many autonomous agents locally or in CI to evaluate workflows and failure modes, including Human-in-the-Loop.

How It's Used

Stress-test agent collaboration patterns and observe failure modes under load
Iterate on multi-agent delegation strategies with parallel runs and configurable retries
Collect interaction logs and outputs to build agent performance histories and reputation signals

Test in the Playground

Validate against scenarios