Cost per month
A team of 5 engineers using Claude Code or Codex versus 1 junior running your factory.
$25k vs $2k a month
Example monthly salaries. AI subscriptions are paid in both setups and not included.
100% free setup
Your own software factory. Request in, shipped feature out.
A line of AI agents that plans, builds, tests, reviews and merges a plain-language request, under rules you wrote down once. We set it up around your product and your release process at no cost to you.
6
stages every request passes through
5
proof checks before any merge
Most small companies have a queue problem, not a software problem
The list of things the product should do grows every week, and the people who can change the code are too expensive to hire, too busy to start, or an agency inbox away.
Hiring a team means months to recruit and a payroll that runs whether or not there is work. An agency turns every change into a brief, an estimate and a wait. Vibe-coding it yourself is fast, but nobody reviews it, nobody tests it, and nothing stops it ignoring how your product is meant to be built.
A software factory keeps the speed of the third option and puts back the discipline of the first.
From a sentence to a merge, in six stages
Intake
A small, fast model sets the type (feature, bug or refactor) and priority. The planner drafts a title, description and acceptance criteria.
Plan
The strongest model splits the work into one to six cohesive tasks, each with the files it may touch. An optional timed gate lets you approve, reject, or let it approve itself.
Build
Tasks run in parallel, each in its own copy of your code. Every task loops build, test and review until its own criteria are met.
Integrate
The branch is rebased on main and your unit, integration and end-to-end suites run. A failure becomes a fix task.
Review
The reviewer marks each acceptance criterion met or unmet, with evidence. Bugs and risky changes can wait at a PR gate for you.
Merge
Checks pass, the change merges, and an optional smoke test runs. If the smoke test fails, the merge is reverted automatically.
Agents write code. Machinery does everything else.
Branching, committing, rebasing, pushing and merging are fixed engine steps with no AI in them. No agent has a tool that can push or merge, and every command an agent asks to run executes in a sealed container with no secrets and no route to your database.
Your rules are enforced, not suggested
A prompt that says “please follow our standards” is a hope. The factory turns your standards into checks the work cannot get past.
- Your conventions, in every prompt. Coding conventions and the lessons from past corrections go into every agent’s instructions. Correct the factory once and the correction sticks.
- Every task stays in its lane. Each task declares the files it may change. Anything written outside them is reverted, and a large stray change stops the work item for you to look at.
- Sensitive code waits for a person. Mark payments, auth or personal data as sensitive. Any change that touches them is labelled high risk and can be held until you approve it.
- No merge without proof. Tests pass, every criterion is met with evidence, scope is respected, the branch is rebased, and your CI is green.
- An audit trail that cannot be edited. Every stage change writes an append-only event: who, what, when and why. Finished work is frozen; reopening creates new work linked to the old.
Human gate: you choose which changes wait for you, and for how long before they approve themselves.
Built around your lifecycle
No two companies ship the same way. During setup we write your process into the factory’s settings, so it is not a document someone has to remember. It is how the machine behaves: which changes wait for you, which paths always raise the risk flag, your install, test and smoke commands, which model plays which role, and the limits on time, retry rounds and parallel tasks.
The factory runs as a set of containers on your own machine or server. Your code and your work history stay with you; only the model calls leave.
One-time setup, then autopilot
Map your product
Repositories, stack, branching, test commands and how a change reaches your users today. Geeks and you.
Write the rulebook
Conventions, sensitive areas, gate policy and who approves what. Your compliance needs go here. Geeks and you.
Fit the line
Agents and models per role, limits, and tests added where your code has none, so the factory has a safety net. Geeks.
First requests together
We run your first real requests alongside you and turn every correction into a lesson the factory keeps. Geeks and you.
Who drives it: someone who knows what the product is for and can tell whether a change works. A founder, product manager, operations or support lead, or your one technical person.
What you bring: your code in a Git repository, a machine or server to run it on, your existing AI subscription (Claude Code, Codex, Cursor or similar), and a person who knows what the product should do.
Same AI, organised instead of improvised
Delivery
Engineers with AI still wait on handoffs, review queues and merge conflicts. The factory runs builders in parallel and moves each work item through every stage without waiting.
Up to 10× the delivery
We compare AI-assisted manual development with an organised factory, not with development without AI.
What a factory does not do
- It does not decide what to build. Choosing what your product does next is still a business decision.
- It is only as safe as your tests. Where coverage is thin, setup adds characterisation tests first, so existing behaviour is pinned down before anything changes.
- Unclear requests come back as questions. It costs you a minute and saves a wrong feature.
- Model usage is a real cost. It is far smaller than a salary and visible per work item.
- Some work items will stop. When rounds or time run out, the item stops and tells you why. Nothing half-finished is merged.
Questions we get
What is a software factory?
A line of AI agents, each with one job, wrapped in machinery that decides what each of them is allowed to do. You write a request in plain words. The factory classifies it, plans it, builds it in parallel, integrates and tests it, reviews it against its acceptance criteria and merges it. Every request becomes a work item on a board, and a card moves only when the factory has finished a stage and proved it.
How is it different from a team using Claude Code or Codex?
The AI is the same. The difference is the human bottleneck. A team using those tools by hand still plans, hands off, reviews and merges manually, and waits on review queues and merge conflicts. The factory runs planning, parallel builders, testing, review and merge as one pipeline against your rules, day and night. The comparison is AI used by hand against AI run as a factory, not against development without AI.
Why is the setup free?
We built the factory for our own products and it is how Geeks Invention ships changes to its internal products today. Setting one up around your product, your rules and your release process is a one-time engineering job that we do at no cost to you. The factory runs on the AI subscription you already pay for, so the running cost is the model usage you already have.
Who drives it day to day?
Someone who knows what the product is for and can tell whether a change works: a founder, product manager, operations or support lead, or your one technical person. They write requests, answer the questions the agents raise, and approve whatever your rules say needs approval. Driving it takes basic technical sense, not deep engineering.
Is it safe to let agents change production code?
The agents write code and nothing else. Branching, committing, rebasing, pushing and merging are fixed engine steps with no AI in them, and every command an agent runs executes in a sealed container with no secrets and no route to your database. Before any merge, five checks must pass: tests, every acceptance criterion met with evidence, scope respected, branch rebased, and your CI green. Paths you mark as sensitive can be held for your approval.
What does it not do?
It does not decide what to build. It is only as safe as your tests, so where coverage is thin setup adds characterisation tests first. Vague requests come back as questions rather than guesses. Model usage is a real cost, visible per work item. And some work items will stop when their rounds or time run out; nothing half-finished is merged.
Related
Get your factory set up. 100% free.
We set it up around your product and your process at no cost to you. It runs on the AI subscription you already pay for.