TNTHUBTrigger & Trial
Review the spec, not the code.
TNTHUB turns a reviewed specification into a running application. The AI drafts the spec, you approve it clause by clause, and the platform builds, verifies, and hands you a live trial environment you can actually use.
Free for individual developers during early access. No card required.
Why specs, not prompts
Prompting a model is a lottery: you cannot review what you cannot see. A specification is a contract. It can be diffed, approved, executed, and held against the implementation that was supposed to satisfy it.
Clause-level approval
Every requirement becomes an individually approvable clause in a Git-diff-style review. You approve intent, not a wall of generated code.
How approval works →
Native spec formats
Gherkin feature files and OpenAPI documents. No proprietary DSL, no translation layer, no lock-in — the tools you already use read them directly.
Explore the spec formats →
Honest completion
Each scenario is graded for assertion strength, so a coverage number cannot be inflated by assertions that pass no matter what.
See assertion grading →
How it works
Three steps. The only one that requires your attention is the one that should.
Draft
The model writes the specification: behavioural scenarios in Gherkin, interface contracts in OpenAPI. Every inference it makes beyond your request is marked for review.
Review & trigger
You work through the spec clause by clause. Once approved, the approved commit becomes an immutable build snapshot and the build is queued.
Verify & trial
The build runs lint, type checks, cross-format consistency checks and the scenario suite, then deploys a disposable trial environment near you with seeded data.
The specification is the contract
You review files your existing toolchain already understands. The same Gherkin that describes a requirement becomes the test that verifies it.
Feature: Refund a settled order
Background: Given an authenticated customer And an order "A-1042" settled 3 days ago
Scenario: Full refund within the window Given the refund window is 30 days When the customer requests a refund for order "A-1042" Then the refund amount is 42.00 EUR And the order status becomes "refunded" And a refund receipt is emailed to the customer
Scenario: Refund is refused after the window Given the refund window is 7 days When the customer requests a refund for order "A-1042" Then the request is rejected with code "REFUND_WINDOW_CLOSED" And the order status remains "settled"</code></pre> <p class="tt-note">The second scenario is the one that matters. A single happy-path assertion would have passed either way — see <a href="/docs/spec-format/assertion-strength/">assertion strength</a>.</p> </div>
How it compares
Stage
Traditional AI coding
TNTHUB
Intent
Buried in a prompt
An approvable clause
Review
Read the generated code
Read a diff of the spec
Correctness
Whatever the model felt like writing
Executed scenarios, graded by assertion strength
Audit trail
Chat history
Approved commit, immutable build snapshot
Delivery
"It should work on your machine"
A deployed trial environment with seeded data
Ownership
Usually retained by the vendor
Your repo, your credentials, your code
Built for contract-clear domains
The platform is strongest where behaviour can be stated precisely and checked mechanically. These are the domains we are tuning for first.
Payments & refunds
Money movement has crisp invariants. A wrong amount is a bug you find in production, not in a code review.
API contracts
An OpenAPI document is already a testable contract. Use it as one.
Risk & trading rules
Decision logic with numeric thresholds is the clearest possible specification.
Device fleets
Command handling, retry semantics and offline behaviour all have exact expected outcomes.
What it does not do
We would rather state the limits plainly than let you discover them later.
Trial is not a sign-off
The trial environment is for experiencing the build, not for formal acceptance. Formal correctness comes from the automated checks — and those are only as good as the assertions in your spec.
It does not write the spec for you
The model drafts it, and the draft still needs a human who knows the domain. If nobody on your team can tell whether a clause is right, no platform can.
It is not a general-purpose code generator
Free-form UI generation and open-ended "build me an app" prompts are out of scope. Spec-shaped work is the target.
It does not take custody of your code
Your spec and your implementation live in your GitHub repository. We are a build and delivery pipeline, not a new owner of your codebase.
Under the hood
GitHub
Specs live in your repository as commits, branches and pull requests. The diff you review is a real Git diff.
Clause-level state
Approval status, verification results and defect attribution are indexed alongside the version they apply to.
Regional trial environments
Trial environments are deployed near you — North America, Europe or Asia-Pacific — so the experience is not a cross-ocean round trip.
Reproducible builds
Every build is pinned to an immutable spec snapshot and a recorded model version, so a result can be reproduced or explained.
Frequently asked questions
Do I need to learn a new specification language?
No. TNTHUB uses Gherkin feature files and OpenAPI documents. If your team already writes BDD scenarios or keeps an OpenAPI document, you are already set up. If not, the template library gets you started in an afternoon.
Who owns the code that gets generated?
You do. It is committed to your repository, with the same licence terms you already operate under. We run the build and the trial environment; we do not take custody of your codebase.
What happens if the model writes a weak test?
This is the central failure mode of automated verification, and we treat it as a first-class problem. Every scenario is graded for assertion strength, weak scenarios are flagged in the completion map, and coverage is always shown next to that grade so it cannot be quietly inflated. Read how the grading works.
Can I use it with an existing codebase?
Yes — the useful unit is the specification, not the repository. Most teams start by writing a spec for one vertical slice, building that slice, and expanding from there.
When does it become available?
We are running a limited early-access programme. Request access and we will get back to you with a slot. We are deliberately not publishing a general availability date we are not certain of.
Start with one vertical slice
Pick one behaviour your team already argues about. Write it as a spec, approve it, and see whether the review is faster than the code review it replaces.