Factory AI launches Environments for Droid coding agents

Factory AI has announced the general availability of Environments across all subscription plans, accessible directly through platform Settings. The feature packages source code repositories, tools, and software dependencies into prebuilt, portable container images for Droids, the company's autonomous coding agents. By separating environment definitions from active runtimes, the platform allows fresh sandboxes to spin up instantly without repeating time-consuming clone, configuration, and setup steps.

To maintain environment integrity and prevent dependency drift, Factory automatically rebuilds and validates these images every few hours, promoting only successful builds to production. This setup underpins scheduled Automations triggered by Slack, GitHub, or webhooks, as well as multi-agent Missions that fan out concurrent subtasks across parallel sandboxes. Factory tracks repository preparation using an Agent Readiness score ranging from 1 to 5. While most enterprise clients initially enter at level 2, Factory notes that multi-agent Missions achieve optimal performance at level 4 or higher.

Enterprise security is managed through centralized administrator policies applied across every managed Droid. Security teams can govern shell command execution by marking commands as allowed, denied, or requiring explicit approval, alongside managing external service integrations called Connectors and packaged capabilities called Skills. Blocked actions are halted at the harness level before execution, while execution hosts can be monitored in the platform's Fleet view. Looking ahead, Factory plans to extend Droid execution capabilities to customer-owned hardware, private clouds, and existing engineering tools.