QM Introduces Multiplayer Agent Harness for Collaborative AI Work in Startups

Multiplayer Agent System for Startups

QM has introduced a multiplayer agent harness specifically developed for startups. This system allows individual employees to have their own isolated workspaces, ensuring independent operation without affecting others. It also facilitates collaboration with AI agents in shared environments such as Slack channels, group messages, and projects.

Scoped Resources and Collaboration

Each user and collaborative room within QM is equipped with its own scoped memory, files, keychain view, permissions, crons, web applications, and a durable sandbox. This design enables personalized agent configurations while maintaining the ability to work collaboratively within shared organizational contexts.

Open Source Approach and Model Flexibility

The platform is built with an open-source philosophy, allowing users to select and switch between various AI models and harnesses. Supported models include Pi, OpenCode, Codex, and Claude Code, ensuring that deployments are not tied to a single vendor. This flexibility allows organizations to choose the AI technology that best fits their needs.

Administrative Control and Customization

QM provides administrators with controls to set organization-level configurations, define security postures, and manage the availability of different harnesses and models. The platform also supports spinning up custom internal web applications and publishing them to specific users, alongside shared skills that can be granted, promoted, and imported from Git repositories.