I have code
- You provide
- Read-only access to the repository.
- You receive
- A report on five areas from an automated scan: secrets, dependencies, exposed AI tooling and default settings, spending caps, logging and alerts.

A written verdict per area from an automated scan, checked by an engineer, after a 15-minute call with a CTO.
TWO TRACKS
The repository track is for code you already have; the specs track, for a project that is still a document. The form asks which one.

REQUEST
No access and no document yet: both come after the call.
Choose the track and tell us what you have. A CTO writes to you within 24 hours with times for the 15-minute call.
To ask something first, write on WhatsApp.
A solid, committed technical team with a great ability to adapt to the needs of the project. […] They integrated very naturally into our team, making collaboration easy from day one.
WHAT WE CHECK
The repository areas are the five failures we keep finding in apps built fast with AI tools. The specification areas are the questions a team needs answered before it can quote and build.
API keys and passwords in the git history, the built frontend, source maps and environment variables shipped to the browser.
Prevents A stolen key paying for someone else's traffic, or opening your database.
Known vulnerabilities in the dependency lockfiles, and security updates a bot opened and nobody merged.
Prevents An exploit built from a public patch reaching an app that never applied it.
Services bound to a public port, default credentials, debug flags, open CORS, service keys in client code and tables without access rules, as far as the code shows them.
Prevents A stranger logging in with the password the install shipped with.
Budgets, rate limits and token limits on model and cloud calls, as set in the code.
Prevents A leaked key or a runaway loop turning into a bill you find at the end of the month.
Error tracking, logs, health checks and alerts, as defined in the code.
Prevents Hearing about an outage or an attack from a customer, or from the invoice.
Whether a team could build from the document: what each feature does, for whom, and how you will know it is done.
Prevents A quote built on guesses, and a build that argues about what was meant.
Which personal or sensitive data the product will hold, where it will live and who can see it.
Prevents A privacy requirement discovered after the data model is built.
Every third party the product depends on, such as payments, an ERP, identity or a model provider, and who owns each account.
Prevents A launch blocked by an integration nobody priced, or an account nobody controls.
Where the project sits in the calculator's range, and what will drive its running cost once it has users.
Prevents A budget that covers the build and not the first year of running it.
Who runs the product after launch, and how they will know it has broken.
Prevents A product that goes live with nobody watching it.
WHAT YOU RECEIVE
Each area gets one of three levels, or is marked as outside the scan's view. The engineer writes the most serious finding in one sentence, with the file, the line and the commit (on the specs track, the section of the document).

| Area | Level | Finding | Evidence |
|---|---|---|---|
| Keys in the wrong place | Act now | A model provider key is in a built frontend file committed to the repository and in two earlier commits; revoke it and issue a new one before anything else. | dist/assets/index-●●●●.js; src/lib/ai.ts, line 12; commits 3f9c… and a71e…; key shown as «xk_l…» |
| Security updates nobody merged | Plan a fix | Four security updates opened by the dependency bot have waited more than 30 days. | package-lock.json; 4 open pull requests, the oldest 47 days old |
| Spending with no ceiling | No finding | A daily budget and a per-user rate limit are set on every model call. | src/server/limits.ts, lines 8–31 |
| Nobody told when it breaks | Outside the scan's view | The alert rules live in the hosting console, which the scan does not see; the report says how to check them. | No alert configuration in the repository |
Every report ends with what the scan could not judge and what the Blueprint would settle, and states that cloud consoles, production configuration and real spend are outside its view.
HOW IT WORKS
Two calls and one report. Nothing is billed at any step.
Choose the track and say what you have. A CTO writes to you within 24 hours with times for the call.
The CTO checks that the diagnostic fits: whose code it is, whether the product is live or planned, and what you want to know. If it does not fit, the CTO tells you on the call.
Repository track: invite a Conectia user with read-only rights, or send a scoped token that expires. Specs track: send the specification.
Within three working days of the day you grant access or we receive the specification. The scan is automated, and an engineer checks the report before it is sent.
A CTO goes through each area with you. Where the findings call for more than a fix list, the CTO offers the Blueprint; the report is yours either way.
ACCESS AND CONFIDENTIALITY
The diagnostic works from the code or the document you give us, never from your running systems.

WHO IT IS FOR
The 15-minute call applies these rules, so you can check them before you write.
WHAT IT IS NOT
It shows what automated checks of the code or the document can see. A penetration test or a certification is a different product, and «No finding» means only that the scan found nothing in that area.
A verdict per area from an automated scan, checked by an engineer. The repository or the specification only, within three working days.
A CTO's written verdict, module by module, with a plan and a closed quote. Two weeks on the code, data and cloud of a running product, or on a specification or a PDF draft.

FAQ
Nothing. The first call, the report and the call to go through it are not billed, and nothing is owed afterwards. The diagnostic comes after a 15-minute call with a CTO, who checks that it fits your case. Where the findings call for it, the CTO offers the Blueprint, and you can decline.
Read-only access to the repository and nothing else: invite a Conectia user, or send a scoped access token that expires. We never ask for credentials by email or chat, never connect to cloud consoles, logs or production, and never run your app. The clone is deleted seven days after the report.
Within three working days of the day you grant access or send the specification. The 15-minute call comes first: a CTO writes within 24 hours of your request with times for it. The call to go through the report follows once you have it.
An automated scan reads it; an engineer checks what the scan reports and writes the most serious finding in one sentence before the report is sent. Nobody reads the code module by module: that is the Blueprint, in which a CTO spends two weeks on the code, the data and the cloud and writes a plan with a closed quote.
No. It is an automated scan of the repository or of the specification; penetration tests and certifications are separate products. It cannot see cloud consoles, production configuration or real spend, and every report says so. «No finding» means the scan found nothing in that area; whether the area is secure is beyond what a scan can say.
Choose the specs track and send what you have in writing: a brief, a PRD or a list of requirements. The report covers scope and acceptance, data, integrations, cost drivers, and launch and operations; AI models under Conectia's own contracts read the document, which is never used to train them. If nothing is written yet, request it anyway: the 15-minute call starts from what you have, your calculator selection included, and the report follows once there is a written specification.
A CTO goes through it with you on a call. Where the findings call for more than a fix list, the CTO offers the Blueprint: two weeks, a plan and a closed quote for the next phase, on your code or on your specification. The report is yours either way, and your own team can act on it.
Because it is how a Blueprint usually starts: the report puts concrete findings from your repository or your specification in front of you, and on the call that goes through it the CTO offers the Blueprint where the findings call for it. The scan is automated, which keeps our cost low enough to give it after the call.