API Documentation

Documentation status

  • Validation: IN PROGRESS – This page summarizes the main APIs and integrations exposed by Qubinode Navigator.
  • Last reviewed: 2025-11-21
  • Community: If you rely on a specific API or discover missing details, please help improve this page via Contributing to docs.

This page is an index of the main APIs and integration surfaces used with Qubinode Navigator. Detailed behavior and examples are documented in the linked pages.

1. Infrastructure Deployment Entry Points

These are the recommended production entry points:

  • ./deploy-qubinode-with-airflow.sh – Unified deployment for Qubinode Navigator plus Apache Airflow and nginx reverse proxy.
  • ./deploy-fastmcp-production.sh – Deploys MCP services only (Airflow MCP server + AI Assistant) using podman-compose --profile mcp.

Platform-specific preconfiguration (inventory, notouch.env, /tmp/config.yml) is documented under ../deployments/.

Developer / legacy scripts

The following scripts are primarily for developers and historical workflows (now located under scripts/development/):

  • scripts/development/setup_modernized.sh – Modern plugin-based setup script for the core Navigator.
  • scripts/development/setup.sh – Legacy setup script (compatibility mode).
  • scripts/development/deploy-qubinode.sh – Older one-shot Qubinode deployment script (hypervisor + Navigator, no Airflow/nginx stack).

See also:

2. AI Assistant REST API

The AI Assistant container exposes a REST API used for health checks and AI-powered operations.

Key characteristics (see AI Assistant Deployment Strategy):

  • Default image: quay.io/takinosh/qubinode-ai-assistant:latest.
  • Default port: 8080 (container) – typically mapped to a host port.
  • Health endpoint:
curl http://localhost:8080/health

Additional endpoints and request/response formats are defined in the AI Assistant service code and may evolve over time. For implementation details, refer to the ai-assistant/ sources and deployment strategy document above.

3. MCP Servers (Model Context Protocol)

Qubinode Navigator ships with MCP servers to expose infrastructure and documentation tools to LLM clients.

High-level documentation:

These cover:

  • Airflow MCP Server – tools for DAG management and VM operations.
  • AI Assistant MCP Server – tools for RAG-powered doc search and chat.
  • Configuration for Claude Desktop and other MCP-compatible clients.

4. Apache Airflow Integration API

The project integrates with Apache Airflow 2.x using Airflow’s standard REST API and DAG interfaces.

Reference docs:

From an API perspective:

  • Airflow’s own REST API is used for:
    • Triggering DAG runs.
    • Inspecting run status.
    • Listing available workflows.
  • Qubinode-specific examples and code snippets are provided in the integration and workflow guides above. For full endpoint semantics, see the official Airflow API documentation.

5. CLI and Plugin Framework

The plugin framework CLI is the main command-line API for interacting with Qubinode Navigator plugins:

# List available plugins
python3 qubinode_cli.py list

# Deploy with a specific OS plugin
python3 qubinode_cli.py deploy --plugin rhel10

# Get plugin information
python3 qubinode_cli.py info --plugin centos_stream10

Additional CLI behavior and plugin options are documented throughout the deployment and development guides. Future iterations may add a dedicated CLI reference page under docs/reference/.

If you need deeper, per-endpoint details for any of the above, please open an issue or contribute examples to this page so it can evolve alongside the platform.

Response:

{
  "field1": "value",
  "field2": 123
}

POST /api/resource

Creates…

Configuration Options

Option Type Default Description
option1 string “default” Description of option1
option2 boolean false Description of option2

Error Codes

Code Description Resolution
E001 Error description How to fix
E002 Error description How to fix