AI can build you an app in an hour.
Then you change one thing, and it falls down.
CodeMySpec is for the morning after. Build software that stays reliable without writing code and without prompting the model. You work through a process instead: say what you want, agree what it should do, and let the harness hold the machine to it.
You don't supervise it. The harness supervises it. You bring your own AI account, nothing traps your work, and building is free.
what a day looks like
A workday with a beginning and an end.
You take in five stories in the morning, it builds all day, you try it in the evening, and it fixes what you found overnight. Then you take the next five. Most projects come out to about ten of these, roughly a business day each.
Take in five stories and answer the design questions.
It builds. You are not in the room for this part.
You click around like a real person and say what's wrong.
It fixes what you found. Come back to a working app.
Start smaller than you want to. Five stories, not fifty — a core you can hold in your head and get running in a day. That limit is for your benefit, not the machine's. If you already drew the whole thing, that work isn't wasted; it just isn't the first move.
the core innovation
If your application is the castle, the BDD specs are the walls.
Every rule you agree to becomes a test that uses your application the way a person would. Given someone who has never signed up, when they sign up, the terms appear. The test clicks the actual button and checks for the actual words on the actual screen. So the machine can't write itself a fake test that passes and move on. It has to make the thing work the way you would find it working.
Because they run, they can't quietly drift from the code, and they can't rot into documentation nobody trusts.
how it works
You work the process. It does the labor.
Five steps, the same five every time, and only two of them are yours. You say what you want and whether it came out right. Everything between those two is the machine's problem, and the harness is what holds it to the job.
no lock-in
It doesn't sit between you and the model. It sits around it.
The harness is a local server on your machine, and each plugin is a thin adapter that points your agent at it. Your requests to the model are never proxied, metered, or marked up. That is also why supporting a new agent is a small adapter rather than a rewrite.
proving it works
"Tests pass" is not the same as "it works."
A QA agent opens your running application in a real browser and uses it — story by story, then all the way through, the way a customer would. That catches the things tests never see: the flow that breaks at the seams, the screen that looks right in the code and wrong to a person.
what it costs
Building is free. You pay when it's running.
Building costs nothing, and that is not a trial. The hosted layer for running and growing the software afterwards — customer email on your own domain, live chat, content publishing, working as a team — is $100 per month per user. Having it built with you starts at $500, quoted. You never pay us for AI usage.
Here is why that split matters, and it is the honest reason to trust this over the alternative. A tool that charges you to build has a reason to keep you building: close enough to the answer to keep paying, never close enough to walk away with a finished thing. Their goal is to keep you subscribed. Your goal is to end up owning something that makes you money or saves you time. We only get paid once your software is live and earning, so getting you out the door with a working application is the entire business.
questions
Straight answers.
What is CodeMySpec?
CodeMySpec lets you build a reliable, maintainable application without writing code and without prompting the model. You work through a process instead: it interviews you about what you want, writes down the behavior you agree on, turns that into tests that check your app the way a person would, and builds until those tests hold. You don't supervise it — the harness supervises it.
What makes CodeMySpec's specifications different?
They execute. CodeMySpec converts a story's acceptance criteria into named rules, Gherkin scenarios, and executable specifications that run against your code, so the behavior you agreed on is continuously checked against what the agent actually built. A written specification is a claim about intent; an executable one is a claim that gets checked. If your application is the castle, these are the walls.
Do I have to write markdown specification documents?
No. The written module specification is a configuration knob you can turn off, and plenty of projects run without it because generating prose documents costs tokens and time. The executable specifications are the part that is not optional, because they are the layer that verifies itself.
Does CodeMySpec replace Cursor, Claude Code, or Copilot?
No — CodeMySpec runs inside them. Those are the coding agent that writes the code; CodeMySpec is the harness on top that adds specs, architecture enforcement, verification, and QA. The tools CodeMySpec is genuinely an alternative to are the app builders like Lovable and Replit: it is the engineer's version of that promise, where the result is code you can still maintain.
What AI agents does CodeMySpec work with?
CodeMySpec works with Claude Code, OpenAI Codex, and Google Antigravity. Claude Code has the deepest integration and is the flagship; Codex and Antigravity are early. It works across agents because the harness is not the plugin — it is a local server on your machine, and each plugin is a thin adapter that points your agent at it.
Does CodeMySpec only work with one kind of app?
No. The core lifecycle — planning the work, agreeing the behavior, writing the tests, and QAing the running app — works on any technology through your AI agent. It goes deepest on the stack it was built for, where it understands sixteen kinds of building block and the rules each one follows.
Does CodeMySpec lock me in?
No. Your executable specifications and your tests are ordinary test files in your own repository, and you bring your own model and API keys with no markup on tokens. There is no proprietary spec format and nothing to export. If you stopped using CodeMySpec tomorrow, you would keep every artifact of value, and the specifications would still run.
How does CodeMySpec catch bugs that tests miss?
A QA agent drives your running application in a real browser — per story, per user journey, or across the whole app — and records an audited trail of every attempt. Passing tests are not the same as working software: unit tests miss integration failures, broken flows, and UI regressions that look correct in the code and break in front of a user.
What does CodeMySpec verify before code ships?
Every time the agent stops, the work is built and run through every check and test before it is allowed to continue, and a failure can stop it there. How strict each check is, is up to you. The point is that the agent cannot quietly leave things broken or skip the tests and tell you it was fine.
What does CodeMySpec cost?
Building is free. The harness and the local server cost nothing to run with your own AI keys, and CodeMySpec never marks up model tokens. The hosted Operate and Grow layer is $100 per month per user, and a done-with-you build engagement starts at $500, quoted.
Who is CodeMySpec for?
Anyone who needs the software to still work next month. That includes people who are not engineers: if you have built something with AI and watched it break every time you change it, this is the layer that was missing. It is not for throwaway prototypes. If you are throwing it away, prompt directly; if you are shipping it, you need the harness.
Try it on one real feature. Building is free.
Install the plugin, run /init to link your project,
then /next to start the next requirement.