Machbase NEO V8.5 Launch and DBMS Overview
Machbase NEO V8.5 ushers in the era of AI agents
Is your data piling up while you struggle to use it to automate your work?
Until now, database management systems (DBMSs) have followed a human-centered model: people had to write SQL and explore data manually before anything could happen.
That era is changing. As AI agents become the users of databases, Machbase is at the center of this transformation.
At 10 a.m. on April 27, 2026, Machbase NEO V8.5 officially launches: next-generation infrastructure designed for AI agents to understand and work autonomously, transforming the way AI data is managed.
Starting today, April 20, we will reveal NEO V8.5’s key features and technologies over the course of a week.
See how we are building the next-generation data platform that AI applications need!
Day 1: Product introduction and NEO DBMS features
SDKs (Node.js/Python/Go/C#), Scalability, and JSON improvements
*Click the image to watch the corresponding video.
Why have conventional DBMSs been difficult for AI agents to use?
AI agents do not click through menus or interpret complex structures the way people do.
They need to understand data immediately, generate queries themselves, and use the results to decide what to do next.
Yet the problems shown below have prevented AI in conventional environments from turning its capabilities into real work.
“Agent-First”: Designed for agents as well as people
Machbase NEO V8.5 is more than a database that answers questions. It is a data platform on which AI agents can perform work directly.
With conventional DBMSs, people wrote queries, interpreted structures, and decided what to execute.
With NEO V8.5, agents automatically generate queries, process structured responses immediately, execute actions, and retry after errors—proactively analyzing data and responding in real time.
▶ Going beyond storage to become the “brain” of AI agents, Machbase NEO V8.5 helps reduce application development costs while maximizing the value of your data.
Why now, and why NEO V8.5?
As AI transformation accelerates, databases are becoming essential infrastructure for entire AI workflows, extending well beyond storage.
Even the best model cannot perform at its full potential if the pipeline feeding it real-time data is slow or unreliable.
Machbase has addressed this challenge directly.
“In the AI era, a company’s core competitiveness begins and ends with data.”
Machbase NEO V8.5 puts that philosophy into practice.
Built on Korea’s unique time-series database technology with world-leading performance in TPC benchmarks,
it provides a high-speed foundation for the source data used in AI analysis, applications, and data operations.
👀 NEO V8.5 at a glance
|
What has changed? Three core areas
1. Usability — Bringing complex AI data pipelines together
A single set of integrated management features handles the complex pipeline from data collection and processing to extraction and analysis.
Workflows that previously required multiple tools can now run on the single NEO V8.5 platform, significantly reducing development time and operating costs.
In particular, a GUI LLM agent based on natural language enables non-specialists without SQL knowledge to analyze and visualize data through conversation. This fundamentally lowers the barrier to working with data.
2. Performance — 10× scalability and 8× faster JSON processing
NEO V8.5 delivers up to approximately 10× greater scalability than the previous version.
It maintains peak performance even under demanding conditions, with 500–1,000 clients connected simultaneously and querying tag tables intensively.
JSON extraction and computation are also up to 8× faster, making it well suited to modern AI workflows that use unstructured and semi-structured data.
3. App platform infrastructure — Building an LLM-connected app ecosystem
NEO V8.5 introduces new JavaScript-based app infrastructure.
Developers can use its JS modules to develop, install, modify, and manage Neo apps. Development environments and sample apps are immediately available for integrating major LLMs, including OpenAI GPT-4, Google Gemini, Meta LLaMA, and DeepSeek.
A closer look at new database engine features
1. Expanded native connectivity across languages
From NEO V8.5 onward, the native protocol supports a broad range of language ecosystems.
Node.js: Easy installation through npm
Python: Installation through pip3
Go: A native Go library available through GitHub
C#: A Universal C# Connector library available through NuGet
Developers can communicate directly with Machbase in their preferred language without an additional translation layer, greatly improving productivity.
2. Powerful engine capabilities — Including more than 20 analysis functions
Custom ROLLUP
High-speed regeneration of statistics across multiple columns supports financial applications. Users can create ROLLUPs with their own criteria, greatly increasing the flexibility of time-series aggregation.
Binary type and view support
Use SQL to work intuitively with bit-level sensor data collected from DAQ equipment, without writing separate, complex binary conversion code.
Meta table improvements
Manage data in a tag table’s metadata area as JSON, making complex hierarchical structures easier to design and operate.
Distance units for tag tables
Alongside time-based analysis, high-speed analysis based on distance is now supported. This is especially valuable in autonomous driving, mobility, robotics, and other fields where physical movement data matters.
16 mathematical functions and more than 20 data analysis functions
Built-in support includes 16 mathematical functions and more than 20 data analysis functions, including P05/P10/P90/P95, PERCENTILE, and the approximate APPROX_PERCENTILE. Perform advanced statistical calculations directly in the database without relying on external analysis tools.
JSON ROLLUP
Apply ROLLUP directly to heterogeneous JSON structures with different members. Automatically generated JSON statistics are easy to use without an in-depth understanding of each internal structure, significantly improving the handling of unstructured time-series data.
Unstructured data processing — From video to JSON
1. Video data processing
NEO V8.5 supports sensor fusion of video and sensor data through the Blackbox app.
Extract footage at a specific time and use it immediately in analysis, supporting fault monitoring and investigation at industrial sites that operate over long periods.
2. Enhanced JSON support
Apply ROLLUP directly to JSON to immediately calculate long-term statistics for data structures with varied configurations.
Using JSON in tag table metadata also makes it easier to design and manage complex hierarchies of tag information.
JavaScript-based Neo app infrastructure
One of the most significant changes in NEO V8.5 is its JavaScript-based app platform infrastructure.
Developers can manage the entire Neo app lifecycle—development, installation, modification, and management—in one environment through JS modules.
The following standard apps are ready to install and use in V8.5.
Replication Automated data replication between Machbase NEO instances. Essential for configuring high availability and disaster recovery. |
OPC-UA Quick and easy data collection from OPC-UA servers. Supports direct integration with industrial sites, including smart factories and defense applications. |
Blackbox An app for collecting video and sensor data over extended periods. Advanced sensor fusion for fault monitoring and post-incident analysis. |
Natural-language GUI LLM An AI agent that uses natural language to analyze data, create apps, and produce visualizations without SQL knowledge. Making data accessible to everyone. |
Key application areas
NEO V8.5 is particularly competitive in the following industries and fields, where high-speed time-series data processing is essential.
High-frequency sensor streams from robots and Physical AI, millisecond market data, real-time quality metrics from production lines, and voltage and temperature time series from large battery packs—all can be handled on a single platform and connected to AI pipelines.
Closing thoughts: AI competitiveness starts with data
Machbase NEO V8.5 is more than a version upgrade.
Under the slogan “Database for AI,” it advances performance, usability, and the app ecosystem together, redefining data infrastructure for the era of AI transformation.
From 10× greater scalability and natural-language LLM agents to an open app platform and native connectivity across languages,
all these innovations are brought together in one product.
Machbase NEO v8.5 provides the data foundation your AI applications need.
With its official launch on April 27, 2026, Machbase will work with customers to take their data capabilities in the AI era to the next level.
Want to learn more about NEO V8.5?
At the webinar in the final week of April, Machbase CEO Sungjin Kim will personally answer questions about these features. We look forward to your registration.
#Machbase #NEO #TSDB #TimeSeriesDataBase #TimeSeries #Data #DBMS #Database #AIDatabase #AIDB #PhysicalAI #DataForAI #LLM #NewFeatures #AIEra #AIApplications #AIAgent #Native #LLMAgent #HighSpeedData #TimeSeriesData #ManufacturingData #IoT #SmartFactory #SensorStreams #FinancialData #Replication #OPCUA #GUI #SensorFusion #Rollup #json #MathFunctions #AnalyticalFunctions #AIEcosystem
