Basa Dev introduces GuardIAn

Your organization's AI,
under control.

GuardIAn is the single gateway between your teams, your agents and any AI model. User governance, budgets, security, compliance and audit, in a single layer, without changing a line of code.

Compatible with the OpenAI API · Deploy on your infrastructure or ours
gateway.guardian.internal
query · reports-agent
compresses context · −78% tokens
GUARDIAN GATEWAY
→ budget model · EU-Frankfurt
Users & GovernanceACTIVE
BudgetsACTIVE
SecurityACTIVE
EU ComplianceACTIVE
Smart routingACTIVE
AuditACTIVE
Architecture aligned with
GDPR
EU AI Act
SOC 2-ready
ISO 42001
OpenAI-compatible
Six pillars, one guardian

Everything you need to govern your company's AI.

GuardIAn sits in the middle of every request, from people or from agents, and enforces your policy before a single piece of data reaches the provider.

Users and governance

SSO, user provisioning, groups and keys for people, agents or third-party software, all treated alike.

Active

Budgets

Assign spend per key, user or group. Automatic cutoff or rerouting to a local model when the limit is reached.

Active

Security

Anti-jailbreak, content moderation and secrets detection. Your own layer or a third party's.

Active

Compliance

Off-the-shelf or custom policies, risk classification and human approval, enabled per entity.

Active

Smart routing

Compresses context and picks the right model (local, budget or premium) with automatic fallback.

Active

Audit

Immutable, chained log, filter-level traceability and exportable, audit-ready reports.

Active
Compliance, piece by piece

Meet GDPR and the EU AI Act without building the report by hand.

  • 1DPA per provider: contracts registered and traceable.
  • 2Risk classification per use case, with human approval for high-risk ones.
  • 3Written consent and configurable data retention.
  • 4Exportable reports mapped straight to the articles of the law.
Compliance status87% ready
GDPR
92%
EU AI Act
81%
ISO 42001
74%
Data retention
95%
TIMESTAMPHASHUSERFILTERS APPLIED
2026-07-10 · 09:41:03a3f9…c21e ⛓u-0042PII masked · EU region
2026-07-10 · 09:41:57c21e…f804 ⛓reports-agentanti-jailbreak · block logged
2026-07-10 · 09:42:12f804…9b32 ⛓u-0042response restored · cost logged
Built for teams already using
OpenAIAzure OpenAIGoogle GeminiAnthropicLocal models
Who we are

Basa Dev builds the trust infrastructure for enterprise AI.

GuardIAn is our flagship product: it was born from watching real teams use AI with no control, no logging and no governance. We built it so that governing AI is as simple as using it. Behind it are 30 years managing critical information for more than 1,200 clients.

6governance pillars in a single layer
<1 dayto connect your first team
0code changes to migrate
100%AI provider-agnostic
Basa Dev · GuardIAn PilotLive

Tell us about your case and we'll set up a guided pilot on your own infrastructure, with your current AI providers.

Talk to Basa Dev

Put the guardian in the middle.

A guided pilot, on your infrastructure, with your current providers.

Contact Basa Dev
Contact

Request a GuardIAn pilot.

Tell us about your case: how many teams use AI, with which providers, and what worries you (spend, security or compliance). We'll set up the pilot on your infrastructure.

No mailing lists, no spam. Your message goes straight to the team.