GS 1등급
GS certification, Grade 1
Certificate No. 26-0296
On-prem
On-premise · Air-gap
Installed on your infrastructure
110+
Built-in tools & plugins
Extend via MCP and APIs
Multi-LLM
Model choice
GPT · Claude · Qwen · Private LLM
Capabilities
Everything an enterprise agent needs, in one platform
Choose how you build, where it runs, and what it connects to.
Vibe Coding in DexOn
Describe it, and it becomes an agent or app
Move between natural-language vibe coding and hands-on full coding inside DexOn — implement, run, test, and fix in one flow.
DexOn
Continue on PC, tablet, and phone
A device-interface workspace that uses files and apps on each device within permitted scope.
Learn morePathFinder
Turn existing systems into agent tools
Connect web systems and APIs as tools — login to test and registration — without code.
Learn moreAgent Runtime · Geny
A runtime that carries work forward
Memory, workspace, tools, and schedules carry context and results into the next task.
Learn moreEasyMode
A guided no-code builder
Combine knowledge, prompts, and tools by drag and drop, then test right away in chat.
Learn moreKnowledge · RAG · Ontology
Knowledge search with citations
Connect documents, file storage, and databases; vector search, rerankers, and ontology improve retrieval.
Tools · MCP
Extend with APIs and MCP
Register external APIs as tools; keys and tokens live in separate auth profiles.
How it works
Build, connect, deploy, govern
01
Build
Canvas, EasyMode, or vibe coding
02
Connect
Knowledge, API tools, and MCP
03
Deploy
Share, embed, schedule; apps publish via admin request
04
Operate
Dashboards, run history, monitoring
Security & Governance
Tested through an attacker's eyes, governed as declared
Each AI service is risk-reviewed and approved through an approval route that matches its grade. Execution is recorded in audit logs.
- Approval routes across partner teams, governance, and security officers
- Guard models, PII masking, and keyword filters
- Reasoning (Harness) separated from execution (Sandbox), isolated per agent
- Access controls re-tested with pentesting in an isolated environment
Deployment
Configured for your infrastructure
Installation, models, and external connections are set to match your security policy.
| On-premise | Air-gapped | |
|---|---|---|
| Location | Your data center and GPUs | Internal network isolated from the internet |
| Models | Internal models; external APIs by policy | Internal models and private LLMs |
| Install & update | GitOps pipeline | Installed after an air-gap transfer process |
| External access | Allowed under a separate transfer policy | Internal systems |
Detailed configuration is agreed based on your infrastructure and security policy. Delivery runs transfer → PoC → build → operate.
Customers
Where XGEN is in use
Building an enterprise AI platform and governance foundation for capital finance
Plateer is building an XGEN-based, on-premise enterprise AI platform that enables Company I to apply AI agents and RAG safely to internal capital-finance work. The engagement also establishes one operating and governance framework across agent development, validation, approval, deployment, operations, and audit.
Jeju Bank's internal generative-AI platform
Building an internal generative-AI platform on XGEN with Jeju Bank, then stabilizing it through on-site operations and widening its use across the bank
An XGEN-based AI coding assistant in an air-gapped network
An XGEN-based AI coding assistant deployed on-premise into an air-gapped development environment with strict security requirements
See how it fits, with your own data
Leave a trial request and we will confirm scope and infrastructure before proposing next steps.