What Factory does
Factory (factory.ai) is an enterprise “agent-native” software development platform that orchestrates autonomous AI systems (“Droids”) to execute work across the software development lifecycle—planning and implementation, pull request review, test generation, documentation maintenance, security checks, and incident response—while keeping humans in control through governance and auditability. The platform is designed to operate across the tools enterprises already use (code hosts, issue trackers, chat/collaboration tools, CI, observability/incident systems), and to route tasks to models in a way intended to balance quality, speed, and cost.
From a product standpoint, Factory markets a “Software Factory” deployment pattern: rather than deploying a single coding agent, Factory coordinates a set of specialized Droids for distinct SDLC activities, with organization-level controls (policies, validation gates, spend/data controls, and analytics) so engineering leaders can supervise agent actions and measure effectiveness tied to shipped work. Factory also emphasizes model independence (supporting multiple model providers and deployment contexts, including local/on-prem for security-sensitive teams), and provides surfaces such as CLI, desktop app, IDE integration, and chat/knowledge workflows through its agent sessions.
Strategically, Factory positions itself as moving enterprises from experimentation to production by combining autonomous execution with controls (risk tiers, allow/deny policying, sandboxing, and audit trails) plus analytics intended to connect adoption/usage to downstream delivery outcomes. Recent updates include expanding partner-led deployment (Factory Partner Network), launching new enterprise analytics for agent effectiveness, integrating with Anthropic’s Claude Marketplace for procurement via Anthropic commitments, enabling local execution on NVIDIA DGX Spark with NVIDIA Nemotron, and broadening the platform into an end-to-end “software factory” system (Factory 2.0).