Code Quality
AI QA Agent for GitLab
Runs structured test plans against your application and reports defects with evidence, connected natively to GitLab so it reads and updates projects,…
Code Quality
AI Test Generation
Writes unit and integration tests for code you already have, as reviewable pull requests, not a manual backlog.
Most codebases have areas nobody wants to touch because nothing tests them. AI Test Generation writes unit and integration tests for the code you already have, following your own framework and conventions, and opens them as reviewable pull requests in your normal workflow. Your engineers review and merge what they like, and coverage grows without pulling anyone off feature work.
- Unit and integration tests written for code with thin coverage
- Follows your existing framework and naming conventions
- Delivered as pull requests your engineers review and merge
$2,900
20–30 working days
Optional maintenance $249/mo
What it is
AI Test Generation writes unit and integration tests for code that already exists in your repository but has thin or missing coverage, and opens them as reviewable pull requests in your normal workflow. It follows your own test framework and conventions.
Why businesses buy it
Low test coverage makes every change risky. Engineers move slowly around fragile code, refactors get postponed, and regressions reach customers. Writing tests for old code is the task everyone agrees matters and nobody has time for. AI Test Generation turns that backlog into reviewable pull requests that match your conventions, so coverage climbs steadily and your team can change code with confidence. The Scope Card fixes the repository, connected systems, workflow and trial runs before you order, so you know exactly what will be delivered.
Problems this solves
Older code has few tests, so every change feels risky
The agent writes unit and integration tests for existing code with thin or missing coverage. Each batch arrives as a reviewable pull request in your normal workflow.
Writing tests for old code never reaches the top of the backlog
Test writing happens in the background through one automated workflow. Your developers review and merge the tests instead of writing them from scratch, so coverage grows alongside normal feature work.
Generated tests that ignore your conventions create clean-up work
It follows your own test framework, conventions and coding standards. A trial on 20 pull requests lets you check the style fits before it runs more widely.
A good fit if
You have existing code with thin or missing test coverage, and writing tests for it keeps losing to feature work.
You want generated tests to arrive as ordinary pull requests, reviewed and merged the way any other change is.
You have coding standards or a testing convention already, even informal, that generated tests should follow.
You are comfortable with tests proposed for review, not merged automatically without a person looking at them.
Not the right fit if
You want existing test plans run and defects reported, not new tests written. See AI QA Agent.
You want pull requests reviewed for bugs and standards, not new tests generated. See AI Code Review Agent.
You mainly want dependency versions checked and updated, not test coverage improved. See AI Dependency Update Agent.
You need broader CI/CD and infrastructure work alongside testing. See AI DevOps Agent.
How it works
- You give us scoped access to one repository, your CI/CD configuration and your testing conventions, such as the framework and file structure you already use.
- The agent reads your existing code and identifies files or functions with thin or missing coverage, prioritizing the ones most likely to break silently.
- It writes unit and integration tests that follow your existing framework and naming conventions, not a generic template unrelated to your codebase.
- Each batch of new tests opens as a normal pull request, so your team reviews, edits or rejects them the way any other change is reviewed.
- No test is merged automatically. A human approves the pull request, the same as any change a teammate would open.
- We trial it on 20 pull requests or pipeline runs before launch, so you see the quality of generated tests against your real code first.
What you provide within 72 hours
Your portal lists these the moment you order. Delivery starts the working day after everything is in and your access is checked.
- Scoped repository access, as an app or token, limited to what the agent needs to open pull requests.
- Your CI/CD configuration details, so generated tests run in your existing pipeline.
- Your coding standards or testing conventions, including the framework and file structure you already use.
- Access to your issue tracker or chat tool, such as Jira, Linear or Slack, if you want progress reported there too.
- Admin access and API keys for every connected system, shared by adding our team as a user or creating a key you control, never by email.
- Your business information (company basics, hours and brand voice), shared during intake.
- One decision-maker who can approve which parts of the codebase to prioritize first.
- An API key for your own AI model account (for example OpenAI or Google Gemini), created for this build with a spending limit.
The 72-hour rule. Send everything on your checklist, including access, within 72 hours of ordering so your build starts on time.
Timeline
Start
Intake confirmed, with repository access, testing conventions and priority areas agreed.
Weeks 1–2
Coverage gaps identified and a first batch of tests generated against your framework.
Week 3
Pull-request workflow and, if chosen, tracker or chat reporting connected and tested.
Weeks 3–4
Trialled on 20 real pull requests or pipeline runs, with results reviewed together.
Week 4
Your review and two rounds of changes.
Go-live
Live, with a recorded handoff session, and 30 days of support starts.
| Option | Delivery time | Price |
|---|---|---|
| Standard | 20–30 working days | $2,900 |
| Express | 15–25 working days | $3,400 |
| Rush | 10–20 working days | $3,900 |
What you receive at handover
- The test generation agent connected to one repository
- Generated tests opened as reviewable pull requests in your Git host
- Configuration based on your test framework, conventions and coding standards
- Test report from a trial on 20 pull requests
- Admin guide and recorded 45-minute handoff session
- 30 days of post-launch support
Expected benefits and how to measure them
- Can make changes to older code feel safer.
- Helps your team add coverage without pausing feature work.
- Coverage on targeted modules
- Run your coverage tool on the chosen modules before launch and again each month after.
- Generated tests merged
- Count test pull requests opened by the agent versus those your team merged.
- Regressions in covered areas
- Count bugs reported in the targeted modules for two months before launch and two months after.
Before you start: Run a coverage report and list the modules with the weakest coverage. Note recent bugs in those areas.
These are things to track, not promised results. Outcomes depend on your data, your team and your market.
Running costs
You pay for the build once. Running the agent costs roughly $20 to $200 a month in your own accounts, covering AI usage and CI minutes, and it moves with how much code it is generating tests for each month.
AI usage during the build and testing is billed to your own accounts too, so your AI model and CI relationships stay in your name.
Ask about running costsCustom requirements
If you need a second repository, more than 2 connected systems, or a second generation workflow, add a note at checkout for a written Change Quote. Your order still runs on its normal timeline. Larger or bespoke projects go through the Custom AI request form.
Request a custom versionKeep it running
Monthly maintenance covers monitoring, fixes, model and platform updates, a monthly health report and change hours for work like a new area of the codebase or updated conventions. It starts after the 30 days of support and can be cancelled any month.
$249/mo (the minimum fee, includes 2 change hours)
Optional. Starts after the post-launch support period. Cancel any month.
How maintenance worksHow this compares
| Do it yourself with a SaaS tool | Hire a custom agency | Upmark Labs | |
|---|---|---|---|
| Price | A test-generation SaaS tool charges per seat every month, generating or not | A contractor writing tests by hand bills by the hour | Ours is one fixed price, shown before you buy |
| Convention fit | A test-generation tool often writes tests in its own generic style | A contractor learns your conventions as they go | Ours follows your existing framework and naming conventions from day one |
| Review step | A test-generation tool may commit tests directly without review | A contractor's tests go through your normal review anyway | Ours opens every batch as a pull request for your team to review |
| Delivery | A test-generation tool is instant to sign up but slow to tune well | A contractor engagement usually takes several weeks to ramp up | Ours has a fixed delivery time shown before you buy, with faster options |
Questions about this product
How is AI Test Generation different from AI QA Agent?
AI QA Agent runs an existing test plan against your running application and reports defects. AI Test Generation writes new unit and integration tests for code that currently has thin or missing coverage, and opens them as pull requests. Use Test Generation first if coverage, not defect reporting, is the gap.
Will AI Test Generation match our existing testing framework?
Yes. The agent is built around the testing framework and file conventions you already use, supplied during intake, rather than a generic template. If your team uses a particular naming pattern or folder structure for tests, describe it during intake and generated tests will follow it.
Are AI Test Generation's tests merged automatically?
No. Every batch opens as a normal pull request, reviewed the same way any teammate's change would be. Nothing merges without a person approving it, which is a fixed part of the build, not a setting you can turn off or need to remember to enable.
How much does AI Test Generation cost to run each month?
Roughly $20 to $200 a month in your own accounts, covering AI usage and CI minutes, and it moves with how much of your codebase is being covered each month. The build itself is paid once and covers connecting your repository and the first coverage pass.
How does AI Test Generation decide which code to test first?
It prioritizes files and functions with thin or missing coverage that look most likely to break silently, based on your codebase and the priority areas you name during intake. You can redirect that priority at any time using the monthly change hours in maintenance.
How long until AI Test Generation is opening pull requests?
Delivery runs on a fixed schedule from a complete intake, and the delivery panel shows that date plus express and rush options before you buy. Most of the time goes into identifying coverage gaps, building tests against your framework, and trialing it on 20 real pull requests before launch.
Coverage gaps closed as pull requests, not a backlog.
Order AI Test Generation and it goes live on a fixed delivery date, opening reviewable pull requests that add tests to your existing code.
Works well with
Code Quality
AI QA Agent for GitHub
Runs structured test plans against your application and reports defects with evidence, connected natively to GitHub so it reads and updates repositories,…
Code Quality
AI Code Review Agent for GitLab
Reviews pull requests for bugs, security issues and standards and comments inline, connected natively to GitLab so it reads and updates projects,…