Factory positions itself as the autonomy stack for enterprise teams and calls its agents Droids, and I want to set the branding aside because underneath it is a competent coding agent platform competing in the most contested category in developer tooling. The things that actually distinguish it are less exciting than the marketing and more useful. First, the pricing is published, in full, across six tiers, which after spending a week on vendor sites that hide behind contact sales is genuinely appreciated. Pro at $20 monthly, Plus at $100, Max at $200, Teams at $60 base plus $40 per seat for up to ten people, then Business and Enterprise on custom terms.
You can work out what this costs before you talk to anyone, and that alone puts it ahead of much of the field on buyer respect. Second, it meets you where you work. Desktop application, CLI and SDK, rather than an editor extension you must adopt wholesale. The SDK matters more than it sounds, because it is what lets you put agent steps inside your own automation rather than only driving them by hand, and teams that get real value out of these tools are almost always the ones that wired them into existing workflow rather than adding another window to alt-tab through.
Third, cloud and local execution are both real options. Droid Computers give you cloud-based agents from the Plus tier up, and local execution stays available, which matters for anyone whose code cannot leave the building. Combined with SSO, SAML and SCIM provisioning and audit logging on the Business tier, this is a product that has clearly been sold into organisations with a security review, and it shows in a good way. Now my reservations.
The tier structure is mostly about usage limits rather than capability. Plus gives you roughly five times Pro's limits, Max roughly ten times, with early feature access thrown in. That is honest, and it also means the question is not which tier has the features you need but how much you will use it, which is very hard to estimate before you start. My advice is to start at Pro, run it for a month, look at where you hit the ceiling, and upgrade on evidence rather than optimism.
Paying $200 a month speculatively for headroom you might not use is a common and avoidable mistake, particularly when several strong competitors sit at or below the $20 mark. On the Droid branding, it is fine and it is also the kind of thing that makes evaluation harder, because every vendor in this space has invented a noun and comparing across them requires translating back to what each actually does. Underneath, this is an agent that reads your repository, makes changes, runs things and produces pull requests, like its competitors, differentiated on enterprise readiness and execution options rather than on some categorically different capability. That is a perfectly respectable position and I would rather they said it.
Learning resources are documentation at docs.factory.ai, plus case studies, a news section and an open source area. Factory for Science is an interesting specialised programme worth a look if you are in research computing. What does not exist is anything teaching agent-assisted development as a practice, which remains the gap across this entire category. Every vendor documents their API and none of them teach you how to decompose work for an agent, how to review generated code without rubber-stamping it, or how to know when to stop delegating.
That skill is the actual determinant of whether these tools help, and everyone is leaving it to folklore. Three point eight. Transparent pricing, solid enterprise engineering, unremarkable differentiation and no teaching.