Native Mojo
The foundation
One foundation under everything we build.
Native Mojo Core is the shared technology foundation every company built here runs on. It is not a product for sale — it is accumulated engineering investment. It exists because the hard parts of business software — security, identity, data, knowledge, events, AI — are the same hard parts every time, and a system that shares them behaves better than six systems that each solved them separately.
Architecture
How the layers relate
Three layers, one direction of dependency. Every company above the core uses the core; nothing in the core knows about what is above it.
Independent businesses
MojoVerify, MojoWare, Mojo Pulse, SitesMojo, MojoWallet — and whatever is built next.
Coordination
People, projects, workflows, knowledge and AI agents working against the same source of truth.
The shared foundation
Security, identity, data, knowledge, AI, events, automation, publishing, open APIs.
Native Mojo Core
The services nobody should have to write twice.
Grouped four ways. Every one of them is a platform service with a documented API, not a library each project copies and drifts from.
Security & identity
Fail-closed permissions
Every endpoint enforces permissions declaratively. An endpoint that forgets to say who may call it does not default to open — it defaults to closed.
Tenant isolation
Organization scoping is a property of the data model, enforced by the framework on read and write, rather than a condition each handler is trusted to remember.
Encrypted secrets
Credentials and keys are stored encrypted at the field level, with envelope encryption backed by a managed key service where the deployment calls for it.
Complete auth stack
Email and passkeys, single sign-on, API keys, service accounts, session policy and multi-factor — with hosted login pages that can be themed per organization.
Structured audit trail
Meaningful changes are recorded as events with the actor, the object and the before-and-after — human or agent, same record.
Data & knowledge
Model-first REST
The API surface is generated from the data model: serialization shapes, search, filtering, pagination and permissions are declared on the model and apply everywhere.
Organizations and membership
Organizations, projects, members, roles and invitations are core concepts — the shape almost every business application needs on its first day.
Organizational knowledge
Documents and decisions are chunked and indexed as vector embeddings so that retrieval is semantic, scoped to the organization, and shared by people and agents alike.
Advanced media storage
Direct-to-storage uploads that never pass through the application, per-organization quotas, generated thumbnails, and URLs that survive redeploys and rollbacks.
Documentation as data
Handbooks and docs live in the platform, publish to sites, and index into the same knowledge store the assistants read.
AI
Assistants inside the application
An assistant embedded in the business application, scoped to that organization's data and permissions, answering from what the organization actually knows.
Agents with real hands
Agents are typed, given explicit tools, and run on pluggable runtimes. They act through the same permitted APIs a person would use — and leave the same audit trail.
Tools, skills, agent types
Atomic tools compose into skills; skills compose into agent types. A workflow that works is captured and reused rather than re-prompted.
Open protocol access
The platform speaks MCP, so the AI tools your team already uses can work against real projects instead of a pasted summary of them.
Operations & publishing
Event & incident intelligence
Events from any source, correlated by rules into incidents, routed to handlers. Detection is the input; the incident is the product.
Asynchronous work
Background jobs, scheduling, retries and dead-letter handling as infrastructure — plus outbound webhooks so other systems can react.
Realtime
Publish and subscribe across processes, so a change in one place reaches every connected client and worker without polling.
Metrics and observability
Time-series metrics and request-level tracing built in, so behaviour under load is a question with an answer.
Publishing infrastructure
Versioned static hosting, custom domains with automatic HTTPS, forms with delivery, media, membership gating and traffic analytics.
Communications
Transactional email with templates and domain verification, SMS, and in-product notification — one path out to people, wherever they are.
Security operations
Detection is cheap. Correlation is the product.
Anything can emit an alert. The value is in deciding which alerts are the same event, which sequence of them is an incident, and what should happen as a result — so the platform models that pipeline directly.
Sensors
Host intrusion detection, application events, infrastructure signals and third-party alerts.
Event
Everything normalizes into one event shape with its source, its subject and its severity.
Rule engine
Rules match, group and threshold events — the same signal five times in a minute is one thing, not five.
Incident
A correlated incident with state, ownership, priority and a history that survives the shift change.
Handlers
Notify, escalate, run an agent, call a webhook, or take an automated action — per rule.
Host intrusion detection is deployed from our own open-source installer, and the correlation layer is a core service — which means an application built on the platform gets security operations without standing up a second system for it.
Open source
The floor is public.
The frameworks the core is built on are open source and installable by anyone. We use exactly what we publish — there is no internal fork with the good parts in it.
The backend framework
Python · Django
Model-first REST, permissions, organizations, jobs, realtime, media, metrics, knowledge and the AI layer.
pip install django-mojo
The frontend framework
JavaScript
The component and data layer our portals are built with, against the same API contracts.
npm install web-mojo
Host intrusion detection
OSSEC
Our packaging and installer for host-based detection, feeding the platform's event pipeline.
github.com/NativeMojoNext
See what stands on it.
Maestro coordinates the work. The ventures are the proof. Both are built on the core described above.