Mimir Foundation

A local-first foundation for evidence-backed intelligence

Mimir provides the platform foundation behind ODIN and DORI.

It is responsible for the backend capabilities that make operational intelligence useful: asset records, evidence references, health and readiness signals, memory and knowledge context, provider routing, governance boundaries, degraded-mode behavior, and capability-pack structure.

Mimir should make recommendations traceable. DORI should be able to distinguish facts, evidence, assumptions, missing information, and confidence.

Foundation capabilities

Capability Role
Asset Registry records services, systems, relationships, ownership, and provenance
Evidence Model attaches certificates, backups, monitors, update facts, observations, and attestations
Heimdall / Observability surfaces monitoring, alerting, telemetry freshness, and degraded-state visibility
Verdandi / Recovery supports backup and restore readiness evidence
Munin / Knowledge preserves operational context, memory, and runbook knowledge
Bifrost / Provider Routing controls local/cloud AI provider routes and visibility
Governance Boundaries keeps risky changes behind explicit approval and safe modes

Trust boundary

The private site should not expose live infrastructure details, hostnames, IP addresses, secrets, or deployment internals.

The important point is architectural credibility: Dravenkor is not trying to build a glossy dashboard over vague data. Mimir is the evidence-backed foundation that makes ODIN and DORI useful.

Reviewer questions

  • Is the foundation credible without exposing too much internal detail?