Documentation by product
Each product has its own documentation, with installation, configuration, use and the complete reference.
The Control Plane, the portal, AI agents, models, knowledge and the governance of the platform.
The administration console: clusters, hosts, services, operations, backups and high availability.
The agent installed on every host: enrollment, trust, tasks, logs and upgrades.
Every catalog service: packages, licences, modes, high availability, dependencies and parameters.
Python, TypeScript, Go and Java: installation, authentication, errors, pagination and a reference of every method.
The complete references of the Control Plane and Navin Manager APIs, route by route and field by field.
The skills, plugins, commands, agents and packs shipped with the platform, and how to enable them per project.
Hardening, identities and access, secrets and the IT vault, file scanning, audit and compliance.
High availability, backup and recovery, upgrades, monitoring and incident procedures.
From installation to production
The path of an operator, in order. Each step says what it does and why, and leads to the detailed procedure.
- Step 1
Prepare
Size the hosts, open the network flows and check the operating systems before any install: this is what avoids failures half way.
Sizing and prerequisites - Step 2
Install the Navin Manager
The administration console is installed first: it then installs, configures and watches everything else.
Manager installation - Step 3
Enroll the hosts
The Navin agent goes on every host and enrolls with the Manager: it runs the tasks and reports the state of the host.
Agent installation - Step 4
Install the services
The wizard picks the hosts, the mode and the bindings, resolves the dependencies and shows every stage of the install.
Services wizard - Step 5
Configure the platform
Create the tenants, connect single sign-on and verify the installation end to end before opening it to users.
First configuration - Step 6
Operate and secure
Monitoring, backups, upgrades, audit and incident procedures: everything needed to run the platform in production.
Operations guides
OS packages and guided installer
DEB, RPM and pacman for the platform, the Navin Manager and the agent, with signed repositories and offline archives.
Identity, access and governance
Keycloak SSO, multi-tenant isolation with Row Level Security, roles, OpenFGA relations and signed OPA policies.
Models and inference
LiteLLM gateway with registries, capacity control, fallbacks and kill switches, managed vLLM and Ollama engines.
File security
Quarantine, then ClamAV and YARA-X scanning of every file before any AI ingestion, failing closed.