Skip to content
INTYGAStart free

INTYGA Intelligence

Know what changed.
See what needs attention.

Bring deployments, monitoring, cloud alerts and authorization activity into one place. INTYGA Intelligence turns connected signals into a focused list of findings, with the evidence you need to decide what to investigate.

  • Open source ยท Apache 2.0
  • Self-hosted
  • Version 1.0

Requires an INTYGA workspace. You run the application and its database as a single instance.

A place to start the investigation.

Your CI tool knows a build failed. Your monitor knows errors increased. Your cloud provider knows an alarm fired. Intelligence brings the signals you connect into a common view and applies explicit detection rules.

Follow a service

Group related deployment, monitoring and operational sources around a service such as your production API. Keep unrelated products and environments separate.

Understand the finding

See measured counts, comparison windows and source references where available. Distinguish a correlated regression from a provider-reported alert.

Record your decision

Mark a finding as investigating, expected or fixed, and leave a note. Route selected findings to Slack or a signed webhook when you need notifications.

Beyond Git. Across the tools you run.

The catalogue includes 21 connectors across code hosting, cloud operations, observability, identity and business systems. Start with one useful signal, then add the sources that give it context.

  • GitHub
  • AWS
  • Google Cloud
  • Azure
  • Grafana
  • Google Workspace
  • Stripe
  • Custom events

Find repeated CI failures, missed jobs, service degradation and unusual refusal patterns. Import selected provider alerts, or investigate error increases after a production deployment. Each integration has a specific scope.

Browse integrations and their coverage โ†’

A routine PR merge is not an anomaly. Changes to sensitive paths appear as review observations unless other evidence supports a finding.

One example ยท Authorization activity

17 refusals. One integration. What changed?

A release service starts attempting permission changes. Its requests are refused, but the volume and action types differ from its observed history. Intelligence surfaces that difference with the evidence attached.

Refusals in 8 minutes
17
Typical for that window
2
Permission-change attempts
12
INTYGA Intelligence example showing a refusal-burst finding with measured counts, a baseline and source references.
Application screenshot with synthetic events generated by the refusal-burst detector. This is an illustrative investigation, not a customer incident.
What does this finding establish?

Refusals exceeded the configured count and baseline thresholds. Twelve attempts involved a permission-change action absent from this subject's earlier observed refusal history.

The finding is a reason to investigate. It does not establish that access was granted or that the service was compromised. No model is needed to produce it.

Follow the full investigation workflow โ†’

You decide what happens next.

Evidence and human review

Detection runs on connected activity. Optional model judgment can add an assessment to eligible findings, but does not originate them or authorize a response. Intelligence does not automatically remediate your systems.

Approval for sensitive changes

Activating sources, changing detection settings and enabling alert destinations require INTYGA approval bound to the change. Passkey acceptance and a successfully applied change are separate states.

Understand data handling and approval โ†’

See where it fits in your stack.

Apache 2.0 application, operated on your infrastructure. Your INTYGA workspace plan is separate. There is no support SLA.

Inspect the source on GitHub โ†’