Your data never leaves
your infrastructure. Ever.

BA Star is deployed inside your environment — your AWS account, your Azure subscription, your GCP project, or your own servers.
We have no access to your data and no visibility into your installation. By design.

Self-hosted means exactly that.

Most AI tools for product teams work like this: you connect your Jira, your Confluence, your documents — and all of that flows through someone else's servers. Their AI processes it, their databases store it, their logs record it.

BA Star works the other way around. You deploy it inside your infrastructure. It connects directly to your Jira, your Confluence, your GitHub — internal connection, internal data. Nothing routes through us.

Your product roadmap, your team's decisions, your requirements documents — these are among the most sensitive assets your organisation holds. They belong on your servers.

What stays with you
  • All your Jira tickets and project data
  • Your Confluence pages and documentation
  • Your codebase and architecture docs
  • Meeting transcripts and requirements docs
  • All generated sessions and artifacts
  • Integration credentials, encrypted at rest

Everything inside your perimeter. One outbound call.

The only data that leaves your infrastructure is the content you choose to analyse —
sent directly to the AI provider. Nothing touches our servers.

Deploy where you already run your infrastructure.

BA Star ships as a single binary, a Docker image, and a Helm chart — the choice is yours. It fits into whatever you're already running.

🐳

Docker & Docker Compose

The fastest way to get started. Pull the image, set your environment variables, run. Works on any Linux host in minutes.

☸️

Helm Chart (Kubernetes)

For teams running AWS EKS, Azure AKS, GCP GKE, or self-managed Kubernetes. Full Helm chart with configurable values.

⚙️

Single Binary

A single compiled binary for Linux (x86_64 and arm64). Drop it on any server, point it at a PostgreSQL database, and start.

AWS / Azure / GCP

Deploy on ECS, App Service, Cloud Run, EC2, or any VM. BA Star has no cloud-provider dependencies — it runs wherever Linux runs.

Air-gapped or on-premise

BA Star can run fully air-gapped. Only the AI analysis calls require outbound internet. If you run a private Claude-compatible endpoint, even those stay internal.

Deployment takes less than an hour.

We walk your technical team through every step — from container setup to Jira integration to first session.

Built for European and enterprise compliance standards.

Because BA Star processes nothing on our servers, the standard GDPR data processor concerns simply don't apply to normal operation. Your data never reaches us.

We provide a Data Processing Agreement covering the narrow case where you might share logs or support information with us directly. But during normal use — generating tickets, indexing your sources, publishing to Jira — there is no data processing on our side.

  • GDPR-compatible architectureYou are both data controller and de facto data processor for your own BA Star installation.
  • No telemetry, no usage trackingWe collect nothing. There are no analytics hooks, no crash reporters, no beacon calls back to us.
  • Credentials encrypted at restIntegration credentials (API keys, tokens) are AES-256 encrypted inside your own database. They never leave your server.
  • OIDC / SSO supportIntegrate with Keycloak, Okta, Auth0, or any OpenID Connect provider — using your own identity infrastructure.

Ready to deploy in your infrastructure?

We'll walk you through the setup and integration configuration.

Get in touch How BA Star works