Does the software work as intended?
A senior engineer checks the code and discusses issues with your team.
Checkevery change
AI governed build
Build your own tools with AI.
Your team builds its own AI tools or software. SevenLab checks security, helps you go live and keeps it running.
For example, a tool for planning, quotes or documents. Your people know the work, our senior engineers check every change.
- 200+
- Projects delivered
- 18
- Years building software
Before a new version goes live, we check the code, data access and security. Then we keep the software running.
Already built something? We check your existing software and help you launch it safely. Your IT team and tools stay, the code stays yours.
Your own people build with AI. Our experienced software engineers check and operate the result. You do not need to hire a full development team for this.
In 60 seconds
Watch the story
How we work
Your team builds. We check and operate.
Start together
We agree what your team will build and set up the workspace and checks. Start with Battlestack or your existing software. Talk to our senior engineers in a shared Slack channel.
way of working agreedBuild with AI
Your team makes its own AI tools or software with Claude Code, Codex or Cursor. A senior engineer checks every proposed code change, called a pull request, before it is merged.
change approvedCheck and launch
We test security, access rights and performance under load. We list the software components used (SBOM) and scan for exposed passwords and keys. A new version goes live only after approval.
approved for launchOperate and improve
SevenLab keeps the software running. We monitor it from the EU, help with incidents and provide a monthly summary of changes, incidents and risks.
monthly summary
Starter kit
Start with Battlestack.
Battlestack is our open-source starting point for your own software. You get a working foundation and rules for your AI tools. The code is on GitHub, in a Git repository you can manage yourself.
npx battlestack@latest my-appcd my-appbattlestack devRequires Node 24 or newer. You only need Docker for the templates with a database.
One command to a running app
You answer a few questions and Battlestack writes a Nuxt 4 app with Docker and CI already set up. Add login and a Postgres database when you need them.
Rules for your AI tools
Every project gets an AGENTS.md and a .mcp.json, so Claude Code and other AI tools start from the same rules.
Stays current
battlestack pull brings improvements into a project you started months ago. battlestack doctor shows where your project has drifted.
Inside a review
How an engineer checks access to data.
Illustrative exampleAI suggestedsupabase/policies/orders.sql
1 -- AI-suggested RLS policy2 create policy "orders_select" on orders3- for select using (true);45 -- Grants every authenticated user6 -- read access to every tenant's orders.
After SevenLab reviewsupabase/policies/orders.sql
1 -- Reviewed RLS policy2 create policy "orders_select" on orders3+ for select using (4+ tenant_id = auth.jwt() ->> 'tenant_id'5+ );6 -- Tenant-scoped, GDPR Art. 32 aligned.
In this example, any signed-in user could see other customers’ orders. The engineer changes the code so users can only read orders from their own organisation.
Checks, launch and operations in one partnership.
Pull request review by a senior engineer
A senior engineer reviews and signs every PR before it is merged.
Shared Slack channel with senior engineers
Direct line to the engineers who review your code. For architecture questions, blockers and pre-PR sanity checks.
Production deployment + IaC
Infrastructure as code, signed releases and reproducible builds. We set up and operate the deployment pipeline.
Audit trail, SBOM and secrets scanning per release
Software bill of materials and secrets scan on every deploy. Audit trail ready by default for ISO 27001; NIS2 and DORA artefacts optional.
EU observability + on-call
Dashboards, alerting and on-call escalation hosted in EU regions, with named engineers responsible for incident response.
Monthly governance report
A summary of releases, incidents, outstanding risks and compliance coverage for your management team.
Onboarding workshop included
We start with a workflow presentation and stack agreement session, set up before the first pull request.
Works with your existing stack
Your IT team stays owner of architecture and tooling, and Claude, Copilot, ChatGPT and Cursor stay in use. We adapt to the stack your team has chosen, with no forced migrations or rebuilds.
Built by SevenLab
200+projects
Delivered for clients across different SevenLab services.
Not a count of AI governed build engagements.
What we review
What we check before your software goes live.
Examples of checks. They reduce risk but do not rule out errors.
Can someone see another customer’s data?
We check who can read and change each set of data.
Accessto the right data
Which software are you using?
We list components and scan for vulnerabilities and exposed keys.
Scanbefore launch
What happens when usage grows?
We test performance under load and set up monitoring and incident support.
Operateafter launch
You know the work.Build your own software for it.
AI lets you create a tool for your daily work. AI governed build means SevenLab checks every change, helps you go live and runs the software.
Coverage
Coverage per framework.
For your IT team and auditors: these checks and documents help explain how your software is secured. Applicable obligations depend on your organisation and use case.
- NIS2 / Cyberbeveiligingswet24h early warningBoard-level accountability, incident reporting within 24 hours for essential entities. optional
- DORAArt. 28 · Art. 17Third-party ICT oversight, audit rights, and major incident reporting for financial entities. optional
- EU AI ActRisk management lifecycleRisk management, data governance and post-market monitoring for high-risk AI systems. covered
- ISO 27001A.8.28 · A.8.30Secure coding controls and outsourced development oversight, evidenced per release. covered
- GDPRArt. 32Appropriate technical measures, including encryption, access control and tested resilience. covered
Coverage means we operate the controls these frameworks require, not that we issue certifications. NIS2 and DORA are not part of the standard engagement. We can set those controls and artefacts up as an option. We work alongside your auditors and provide the artefacts they need.
What SevenLab has already shipped to production
AI governed build is our own offer, not a client case. These are projects SevenLab built and delivered for clients, with measurable outcomes.
AI Agent & Intervention Management
Managing Dutch absence cases with an integrated AI platform
Reduced administrative load by 75%
AI Agent
An AI assistant for gas tanker safety procedures
Reduced procedure lookup time by 90%
AI Training Platform
Digitising safety training for a multilingual industrial workforce
Reduced onboarding time by over 90%
One partnership, from checks to operations.
The starting workshop, security checks, operations and monthly reporting are included.
AI governed build
Your team builds its own AI tools or software with AI. We check and operate it.
From EUR15,000/month
- Every code change checked by a senior engineer
- Direct contact with our engineers through Slack
- Security checks and approval before launch
- EU monitoring, incident support and a monthly report
No sales pressure
Frequently asked questions
Everything you need to know about AI governed build.
See how AI governed build fits your team.
A 15-minute call, no sales pitch. We look at where your team stands and show you what a senior review looks like on your code. Bring your IT lead if you like.
Talk directly with our senior engineers
15 minutes, no obligation
No sales pressure