LAYER 03 · AI APPLICATION RUNTIME
Beyond storage and services,
applications that deliver results in the field
Built-in apps, custom apps, and third-party apps can use data and services directly.
AI Coding Agent and JSH make it easy to extend the system to meet your business needs.
JSH · MACHBASE NEO 8.5.0+
One application runtime
from AI-assisted development to operations
Run `.js` files and directories containing `index.js` as applications.
A virtual filesystem, package installation and service controls bring your applications from development into operation.
Application Structure
Single .js files or app directories with index.js
Service Manager
Manage long-running apps with install, start, stop, and status
Virtual File System
Isolated /work and host directory mounts
Built-in Modules
DB·HTTP·MQTT·NATS·OPC-UA·WebSocket
Package Manager
Manage package.json, dependencies, scripts and GitHub packages
Native Data Access
SQL queries, data ingestion and TQL SCRIPT integration
BUILD
AI Coding Agent
INSTALL
Package Manager
RUN
JSH Runtime
OPERATE
Service Manager
BUILT-IN INDUSTRIAL APPLICATIONS
Use built-in apps immediately after installation and extend them for your operations.
NEO
BUILT-IN APP 01
Data Replication
Select only the raw TAG and LOG data you need for real-time replication across edge devices, servers and the cloud. After an interruption, replication resumes from the last processed position.
Use cases
Reliably consolidate data from multiple production edge systems on a central analytics server
Source·Target Table Mapping
Tag prefixes, suffixes and selective replication
Tag Prefix, Suffix, and Select Duplication
Resume from checkpoints · Monitor gaps
BUILT-IN APP 02
OPC-UA Client
Browse OPC-UA Server nodes and configure the collection interval and storage destination. Separate the Collector and Database Server to scale the system to suit your on-site setup.
Use cases
Collect PLC and equipment data using standard OPC-UA and store it in a TAG Table
Multiple Collector Registration
Tag Name mapping for each node
Register multiple Collectors
Calculate and process collected values, then store them in tables
BUILT-IN APP 03
Blackbox
Manage camera video and sensor data on the same timeline. Jump to an object recognition event to examine video frames and measurements together.
Use cases
Track video and sensor signals together before and after equipment anomaly events
Video saving and real-time viewing
Record object recognition events
Video Panel Time Synchronization
Jump to event times and save navigation logs
BUILT-IN APP 04
LLM Chat
Explore Machbase tables, tags and time-series data with natural-language questions. The LLM Agent creates SQL and TQL, and generates analytics dashboards and HTML reports.
Use cases
Analyze the causes of equipment vibration and generate diagnostic reports without writing SQL
Claude·ChatGPT·Gemini·Ollama
Dashboard · HTML Report
Natural Language SQL and TQL Generation
Analysis Template and Data Collection Timer
OPEN DEMO REPOSITORIES
Validate Physical AI data
in real applications
Explore the installation setup, data loading, queries, and UI implementation in the public repository. Each image shows the corresponding official demo in action.
Robot Arm Simulation
Validate 3D robotic arm replay, timeline, frame query and JSON path ROLLUP on a single screen.
LiDAR Data Simulation
Store KITTI LiDAR point and distance data, and examine changes in 3D space and query latency.
Humanoid Simulation
Collect humanoid joint, posture and state data as time series and compare motion anomalies.
Railway APU Diagnosis
Calculate pressure, startup and load risks from MetroPT-3 APU signals and express them as a health index.
Prometheus Monitoring
Scrape Neo's metric endpoint with Prometheus and monitor operational time series in Grafana.
BUILD YOUR OWN APP
Product data and app code
in a single deployment unit
JSH apps directly use Neo's database and protocol modules. Install packages from GitHub or deploy a custom app directory, then manage the apps with Service Manager.
Package Manager
Service Manager
Machbase Github
See how the three layers connect at a glance.
PRODUCT OVERVIEW
Return to the complete product architecture, from Database Engine to AI Application Runtime.
