Factory Review Australia 2026: Droid Agents for Serious Engineering Teams

Verdict (TL;DR) Factory is an agent-native coding platform built around "Droids." The depth is real. The platform spans CLI, desktop and SDK. It runs frontier and open models, and deploys on-premise or air-gapped. But billing is USD-only. GST treatment is unstated. No Australian data region is named outside Enterprise controls. This is a developer tool, not a tool for most of our readers. Engineering teams should shortlist it. Solo non-coders should not. Score: 15/25 — ★★★ Good Last reviewed: 28 June 2026 by Adam.

Most tools we review are for a busy Melbourne business owner. Factory is not.

Factory builds AI software-development agents it calls Droids. They write, test and ship code. They run in a terminal, a desktop app, or through an SDK. The target buyer is an engineering team, not a marketer with a side project.

We still reviewed it. Australian dev teams are evaluating these agents now, and the local-fit questions — pricing currency, GST, data residency — are exactly where this publication adds value. So this review is honest about who Factory suits, and who it does not.

What Factory actually is

Factory describes itself as "agent-native software development." In plain terms, the Droid is an autonomous coding agent. You give it a task. It plans, edits files, runs commands and proposes changes.

Per Factory's own site, the platform works across the Factory App, the Droid CLI and a Droid SDK. The company positions this as model-independent: the agent can use leading frontier and open-weight models and pick one based on the job (factory.ai homepage; see Sources).

Factory's pricing documentation states that "Droid sessions consume Factory Standard Credits based on model usage, along with compute usage for Droid Computers" (docs.factory.ai/pricing; see Sources). Droid Computers are Factory-managed cloud machines the agent runs on for remote workloads.

The homepage frames Factory as a system for the whole software lifecycle — triage, code generation, validation, release, documentation and monitoring. It lists deployment options including fully managed SaaS, hybrid cloud, on-premise and air-gapped, across AWS, Azure and GCP (factory.ai homepage; see Sources).

That last point matters. Air-gapped and on-premise deployment is rare among coding agents. For a bank, an insurer or a government contractor, it can be the difference between a trial and a hard no.

The multi-model stance is the other standout. Factory's homepage states the platform supports "all leading frontier and open-weight models, including the latest versions of GPT-5, Claude Opus and Sonnet, Gemini, and more." We are reporting that as Factory's own claim about the models it integrates, not as an endorsement from those model vendors. The practical upshot: you are not locked to one lab's pricing or rate limits. If one model degrades or gets expensive, you switch. For a team managing cost across a year, that flexibility is worth real money.

What Factory does not do is hold your hand. There is no drag-and-drop builder, no visual canvas, no "describe your app and watch it appear." You work in a terminal or an SDK, you read diffs, and you approve changes. That is the right design for its buyer. It is the wrong design for most people who land on this site.

Pricing — and the Australian catch

Factory's pricing page lists five tiers. All figures are in US dollars. We convert at roughly 1.40 and flag GST below.

Source: factory.ai/pricing (now factory.com/pricing; re-read in a browser 30 September 2026) and docs.factory.ai/pricing (see Sources).

Pro at US$20 a month is roughly A$28. That is competitive on paper against other terminal agents. Plus is about 5x Pro's usage; Max is about 10x, plus early access to new features — Factory's own wording on both pricing pages.

Here is the catch for buyers here. The usage model is credit-metered with rolling rate limits. Factory's docs describe 5-hour, weekly and monthly rate-limit windows that "run independently." Heavy sessions burn Standard Credits faster. Your real monthly cost is harder to predict than a flat seat price suggests.

Factory also sells Extra Usage: prepaid, pay-as-you-go credits with a US$10 minimum that "never expire," consumed only after Standard Usage runs out (docs.factory.ai/pricing). Individual plans include a "free BYOK usage allowance" for bring-your-own-key model access.

GST and currency. Billing is in USD. Factory's pricing pages do not state Australian GST treatment. Australian businesses should expect to handle GST on imported digital services themselves and budget for card foreign-exchange fees on top of the headline price. We could not verify a published AUD price, so we have not asserted one.

Teams now has a published price: US$60 a month per team plus US$40 a month per seat, for up to 10 seats (a five-person team is about US$260 a month, ~A$364). Business and Enterprise are quote-only. Business covers up to 150 seats with custom usage limits; Enterprise lifts the seat cap and adds dedicated compute. Both are described as not affected by the standard rate-limit changes.

We found no free trial or free tier published on either pricing page. If you want to test Factory at no cost, the BYOK allowance is the closest thing on offer.

Security, compliance and data residency

This is where Factory is stronger than most young agent vendors — with caveats.

Factory's security page states the company does not use customer code as training data: "Your code is your IP... that means not using your code as training data." It lists data encrypted at rest with AES-256 and in transit with TLS 1.2 or higher. It describes hosting "within a sandboxed single-tenant environment" with its own VPC, configurable audit logging and permissions enforcement (factory.ai/security; see Sources).

On certifications, Factory's security page references SOC 2, GDPR, CCPA and ISO 42001 — the AI-management-systems standard. We are attributing those to Factory's own page, not independently auditing them. Some third-party write-ups also claim SOC 2 Type II and ISO 27001; we could not confirm those on Factory's own pages, so we are not stating them as fact.

Data residency. This is the key Australian gap. Factory's pricing page lists data residency as an Enterprise-tier admin control, alongside customer-managed encryption keys and network policy. Zero Data Retention appears at the Business level. Factory does not publish a named Australian hosting region on the pages we reviewed.

So: residency controls exist, but only on the top tier, and no AU region is advertised. A Pro or Plus subscriber here has no published guarantee about where their data sits. For regulated Australian work — finance, health, government — that is a question for Factory's sales team before you commit, not an assumption you can make.

This is general product information, not security advice. Verify Factory's current certifications and residency options directly with the vendor before relying on them.

How it scores

The components sum to 15, matching the headline and the frontmatter.

Australian Fit gate: Australian Fit scores 2/5, below the 4/5 threshold. Factory therefore cannot be named a category Top Pick and cannot wear five stars, regardless of total. At 15/25 the star band is three stars anyway, so the gate does not change the displayed rating here. See our Methodology for how the gate works.

Who should buy it

Buy Factory if you run an Australian engineering team that wants autonomous agents across the full software lifecycle, and you need on-premise or air-gapped deployment that most rivals can't offer. Start on Pro to learn the credit model, then move the team to Teams or Enterprise where rate limits and residency controls actually apply.

Do not buy Factory if you are a non-technical operator, a solo founder without a codebase, or anyone who wanted "AI that builds my app for me" in plain English. That is not this product. For that audience, see our best AI app builders and vibe-coding tools roundup.

Wait and watch if Australian data residency is non-negotiable but you can't reach Enterprise spend. Factory has the controls; it just hasn't published an AU region or AUD pricing. Both could change. When they do, this score moves up.

The verdict

Factory is a genuinely capable agent platform held back, for our readers, by where it points its billing and its hosting. The engineering is not the problem. The local fit is.

For Australian dev teams with real codebases and real compliance needs, it earns a place on the shortlist next to the other terminal agents in our best AI coding tools roundup. For everyone else this publication usually serves, it's the wrong tool — and we'd rather say so plainly than dress it up.

Three stars. Good, not yet built for Australia.

Sources

All claims above are anchored to Factory's own pages. The Wayback snapshots below are the closest verified captures available at the time of writing (16–27 June 2026); a same-day re-snapshot is queued for the pre-publish check.

• Factory pricing (plans, USD prices, Teams/Enterprise, residency-as-Enterprise-control): https://factory.ai/pricing — snapshot: http://web.archive.org/web/20260618223613/https://factory.ai/pricing
• Factory documentation pricing (credits, rate-limit windows, Extra Usage, BYOK, no free tier): https://docs.factory.ai/pricing — snapshot: http://web.archive.org/web/20260616034339/https://docs.factory.ai/pricing
• Factory security (no-training, AES-256, TLS 1.2+, single-tenant VPC, SOC 2 / GDPR / CCPA / ISO 42001, audit logging): https://factory.ai/security — snapshot: http://web.archive.org/web/20260617190547/https://factory.ai/security
• Factory homepage (agent-native, CLI/App/SDK, lifecycle stages, deployment options): https://factory.ai/ — snapshot: http://web.archive.org/web/20260627084620/https://factory.ai/

Exchange rate used: US$1 ≈ A$1.40 (September 2026). GST: Factory bills in USD; Australian GST treatment is not published by the vendor.