Skip to content

Obsidian Core™

V3

Enterprise Fabric

A product can be installed.

Infrastructure has to survive the organization built around it.

V3 is where Obsidian Core™ moves beyond individual deployments. Distribution, multiple sites, extensions, enterprise identity, local model orchestration, integration, failover, and organizational scale become part of the platform itself.

V3 Roadmap

V3 Roadmap

The Core becomes something other systems depend on.

That changes the standard. Installation is no longer enough. Updates have to be controlled. Sites have to communicate. Extensions have to behave. Identity has to scale. Failure has to be planned for.

The first release is one installation for one company, including multiple locations on that installation. V3 is the later architecture when an organization runs more than one Core.

Distribution

Deployment cannot depend on us touching every machine.

V3 moves toward direct online distribution, organization provisioning, verified installation packages, controlled updates, staged rollout, rollback, and version compatibility.

Secure deployment for a single company host is the current direction. V3 changes the scale: signed distribution, controlled updates, and version compatibility across many installations. This page does not offer a public download.

Direct online download

Signed distribution packages with a guided installer and bootstrap path.

Organization provisioning

License provisioning and controlled onboarding without App Store dependency.

Update channel

Verified updates, staged rollout, rollback, and version compatibility checks.

Signed software pipeline

Signature checking and controlled software-update infrastructure.

Multi-site Core

One organization may have more than one Core.

V3 is designed toward multiple authorized Core nodes operating across offices, facilities, entities, or locations while preserving local operation and controlled synchronization.

The difficult part is not connecting them. The difficult part is deciding what moves, when it moves, who can see it, and what happens when systems disagree.

Distributed nodes

Office, site, and facility deployments with an organization-wide node registry.

Selective replication

Encrypted site-to-site replication of selected records between trusted sites.

Cross-site context

Distributed work queues, workload routing, and location-aware services.

High availability

Critical systems need a plan for failure.

V3 expands toward redundant nodes, health monitoring, failover, backup verification, recovery orchestration, and controlled restoration.

Redundant nodes

Failover, health monitoring, and replicated critical services.

Service continuity

Continuity across controlled local clusters—not SaaS-style tenancy.

Backup verification

Recovery verification and controlled restore testing.

Disaster recovery

Orchestrated restoration for larger deployments.

Enterprise identity

Hundreds of users cannot be managed like ten.

V3 moves toward enterprise identity, organizational directories, automated provisioning, federated sign-in, device trust, and policy-based authorization.

Specific vendors are not claimed.

SSO & directories

Organizational identity systems and federated sign-in.

Provisioning

Automated user provisioning and SCIM-style directory flows.

Policy authorization

Policy-based authorization and device trust beyond simple role checks.

Extension platform

A platform that cannot be extended eventually becomes the limitation.

V3 introduces the direction for a documented extension SDK, signed modules, defined permissions, compatibility contracts, publisher identity, and organization-specific extensions.

The Core should be able to grow without turning itself into an uncontrolled plugin folder.

Developer SDK

Documented APIs, module lifecycle, and version compatibility contracts.

Signed extensions

Publisher identity, extension permissions, and plugin sandboxing before load.

Integration gateway

The Core will not be the only system in the building.

V3 is designed to connect with accounting, ERP, CRM, payroll providers, calendars, document systems, databases, payment infrastructure, and internal APIs without giving every outside system unrestricted access to the Core.

Payroll expansion

Enterprise payroll extends the V2 engine.

Multi-state payroll, multi-entity payroll, benefits integration, provider connectors, direct-deposit workflows, tax-filing workflows, compliance feeds, and consolidated reporting sit here as expansion—not as present claims.

Multi-state / multi-entity

Payroll across entities and jurisdictions.

Provider connectors

Benefits, payroll providers, direct-deposit and tax-filing workflow integrations.

Compliance feeds

Update feeds and consolidated reporting for larger organizations.

Boundary

External-network payroll functions are not claimed as shipping.

Model orchestration

One model will not be the best tool for every job.

V3 expands toward multiple local models, workload-specific routing, local inference nodes, GPU scheduling, capability-aware model selection, and controlled model lifecycle management.

The model becomes a component. The Core remains the system.

Multiple local models

Specialized selection by workload without mandatory cloud AI.

Routing

Capability-aware routing, GPU workload direction, and local inference nodes.

Lifecycle

Model version control as part of the fabric.

Organization-wide intelligence

A larger organization creates a larger memory problem.

V3 is designed toward cross-site search, connected organizational knowledge, forecasting, and planning across permitted operational data. Operations Analytics on a single company installation is already part of the first-release direction and is not deferred here.

The objective is not a prettier dashboard. It is a system that can see enough of the organization to be useful.

Cross-site knowledge

Organization-wide permitted search and cross-site knowledge graph work.

Analytics & planning

Operational analytics, trends, forecasting, anomaly detection, and resource utilization.

Engineering

V3 is not a feature list.

It requires distributed systems, replication, software signing, update infrastructure, enterprise identity, extension boundaries, failover, workload scheduling, model orchestration, and integration architecture.

If those problems are not solved properly, the rest does not matter.