← Back to Discover

Factory

Platform for building software with autonomous agents called Droids that plan, build, review and ship code, with model routing, code review, automations and usage analytics aimed at enterprise teams.

Coding agentsCode review

What it does

Factory is a platform for software development with autonomous agents, which it calls Droids. The site describes agents that plan, build, review and ship software, usable across desktop, terminal, web and CLI. Teams set the goals and policies, and Droids carry out and validate the changes. The pitch is control for organizations: they decide where the system runs, how it learns and which models it uses. Model Routing sends each task to the model that best balances quality, speed and cost, and the site describes the platform as model-independent. Around the agents sit tools for oversight and rollout. Analytics tracks adoption and consumption per developer and team, Agent Readiness scores each repository over time, Mission Control follows active missions with execution logs, and Droid Computers are purpose-built hardware that runs agents locally with state kept across sessions. Other parts include Automations for routine engineering work, Code Review with inline comments on every pull request, and AutoWiki, a wiki for each repository that is kept current as the code changes.

Capabilities

  • Droid agents that plan, build, review and ship software
  • Model Routing that picks a model per task by quality, speed and cost
  • Code Review with inline comments on every pull request
  • Automations for routine engineering work
  • Mission Control, Agent Readiness scores and per-team usage analytics
  • AutoWiki: a living wiki for each repository

When to use it

  • An engineering organization wants to roll out agents with governance and usage reporting
  • You want pull-request reviews and routine tasks handled by agents working from shared context
  • You want to choose which models run which tasks instead of depending on a single one
  • You want to measure how ready your repositories are for agent-driven work