GovOn

AX (Agentic Transformation) platform that unifies Korea's fragmented government DX infrastructure through agentic AI.

---

Korea's public sector has achieved widespread DX (Digital Transformation) — but data and APIs remain siloed across ministries, departments, and agencies, each with independent systems. GovOn is the AX (Agentic Transformation) layer on top. It wraps each agency's API endpoints as LLM-callable tools, and a central LLM autonomously decides which tools to call, chains them, and synthesizes results — acting as a single unified interface over the entire government infrastructure.

The CLI is lightweight (~10 MB). The heavy lifting happens on a remote server running EXAONE 4.0-32B with domain-specific LoRA adapters. MVP supports 2 domain adapters (공공행정, 법률); the roadmap targets 16 national domains.

Quick Start

npm install -g govon
export GOVON_RUNTIME_URL=https://umyunsang-govon-runtime.hf.space
govon

Installation

Prerequisites: Node.js 18+

npm (recommended)

npm install -g govon

Native installer

curl -fsSL https://raw.githubusercontent.com/GovOn-Org/GovOn/main/scripts/install.sh | bash

Homebrew (macOS / Linux)

brew install govon-org/tap/govon

Update

# npm
npm install -g govon@latest

# Homebrew
brew upgrade govon

Self-hosted runtime (GPU required)

You can host your own GovOn runtime on Hugging Face Spaces or on-premises.

Option A: Hugging Face Spaces (recommended)

1. Duplicate the reference Space: umyunsang/govon-runtime
2. Select A100 GPU (80 GB) or higher hardware
3. Wait for the Space to start, then set your CLI to point at it:

export GOVON_RUNTIME_URL=https://<your-hf-username>-govon-runtime.hf.space
govon

Tip: Add the export to your shell profile (~/.zshrc, ~/.bashrc) so it persists across sessions.

Option B: Docker (on-premises)

govon server pull
govon server start

Option C: pip extras (bare-metal)

pip install govon[server]

Hardware requirement: The runtime requires an NVIDIA GPU with ≥ 80 GB VRAM (A100 80 GB or higher). Smaller GPUs will fail to load the EXAONE 4.0-32B model.

Usage

Interactive mode (REPL)

# Point the CLI to your runtime (HF Space, Docker, or localhost)
export GOVON_RUNTIME_URL=https://<your-hf-username>-govon-runtime.hf.space
govon

govon> Draft a response for a road damage complaint

┌─ Approval Request ────────────────────┐
│  Type: Draft response                 │
│  Goal: Road damage complaint response │
│  Tasks:                               │
│   - Look up legal basis               │
│   - Search similar cases              │
│                                       │
│  ● Approve  ○ Reject                  │
└───────────────────────────────────────┘

The agent proposes a plan and waits for your approval before executing any tools.

One-shot mode

govon "Show me road damage complaint statistics for this month"

Multi-turn conversations

govon --session my-session

govon> What are the most common complaint types?
→ (response with statistics)

govon> Draft a response for the top one
→ (uses conversation context to generate a draft)

API call

curl -X POST $GOVON_RUNTIME_URL/v3/agent/run \
  -H "Content-Type: application/json" \
  -d '{"query": "Show complaint statistics", "session_id": "demo-1"}'

Features

• AX platform architecture -- government agency APIs wrapped as LLM tools; central LLM orchestrates cross-agency data access through a single interface
• ReAct agent with 7 tools -- the agent autonomously selects and chains tools based on your request
• Human-in-the-loop approval -- every tool execution requires explicit user approval before running
• Multi-turn conversations -- session-based context management with extractive summarization
• Multi-LoRA inference -- domain-specific adapters loaded per-request on a single base model
• Streaming responses -- real-time SSE streaming with per-node progress display
• Three installation methods -- npm, Homebrew, and native installer

Domain Adapters

Adapter	Domain	Status
public_admin_adapter	공공행정 (Public Administration)	Active (MVP)
legal_adapter	법률 (Legal)	Active (MVP)
(14 more)	과학기술, 교육, 교통물류, 국토관리, 농축수산, 문화관광, 보건의료, 사회복지, 산업고용, 식품건강, 재난안전, 재정금융, 통일외교안보, 환경기상	Roadmap

Tools

Tool	Purpose
api_lookup	Query government data APIs
issue_detector	Detect trending issues across domains
stats_lookup	Retrieve domain statistics
keyword_analyzer	Analyze keyword trends
demographics_lookup	Look up regional demographics
public_admin_adapter	Generate public administration response drafts
legal_adapter	Search legal references and precedents

Architecture

graph LR
    subgraph Client
        CLI["govon CLI<br/>npm / brew"]
    end

    subgraph Server ["AX Runtime (HF Space or Docker)"]
        API["FastAPI"]
        AGENT["ReAct Agent<br/>LLM Orchestrator"]
        LLM["EXAONE 4.0-32B<br/>+ domain LoRAs"]
        TOOLS["Domain Adapters<br/>public_admin / legal<br/>(+ 14 roadmap)"]
    end

    subgraph GovInfra ["Government DX Infrastructure"]
        GOV1["Ministry API A"]
        GOV2["Agency API B"]
        GOV3["Department API C"]
    end

    CLI -- "HTTP / SSE" --> API --> AGENT --> LLM
    AGENT --> TOOLS
    TOOLS --> GOV1 & GOV2 & GOV3

    style Client fill:#e0f2fe,stroke:#0284c7
    style Server fill:#fef3c7,stroke:#d97706
    style GovInfra fill:#f0fdf4,stroke:#16a34a

Loading

DX → AX: The CLI is the entry point; the server is the AX orchestration layer.

• CLI: React + Ink + TypeScript TUI (Node.js 18+, no GPU needed)
• AX Runtime: EXAONE 4.0-32B + vLLM + Multi-LoRA on A100 80 GB
• Domain Adapters: Each adapter wraps a government agency's DX APIs as LLM-callable tools

Server Management

Manage the Docker-based backend with govon server:

Command	Description
govon server pull [TAG]	Pull the Docker image
govon server start	Start the backend (docker compose up -d)
govon server stop	Stop the backend (docker compose down)
govon server status	Check container status + /health endpoint
govon server logs	Stream logs in real time

Configuration

Variable	Description	Default
GOVON_RUNTIME_URL	Server URL	http://localhost:7860
API_KEY	API authentication key	(none -- unauthenticated access)
HOST_PORT	Local server port	8000

export GOVON_RUNTIME_URL=https://umyunsang-govon-runtime.hf.space
export API_KEY=your-api-key

See deploy/env/.env.example for the full list of server-side configuration options.

Documentation

Resource	Link
User Guide	docs/guide/user-guide.md
Operations Guide	docs/guide/ops-guide.md
API Reference	Endpoint reference
Demo Scenarios	docs/demo/README.md
Docs Portal	govon-org.github.io/GovOn
Public Roadmap	Workstreams
Community Discussion	GitHub Discussions

Resources

Package	Install
npm	npm install -g govon
Homebrew	brew install govon-org/tap/govon
Native	curl -fsSL .../install.sh | bash
Docker	ghcr.io/govon-org/govon
HF Space (hosted runtime)	umyunsang/govon-runtime
Public Admin LoRA Adapter (public_admin_adapter)	umyunsang/govon-civil-adapter
Legal LoRA Adapter	siwo/govon-legal-adapter
Releases	GitHub Releases

Contributing

git clone https://github.com/GovOn-Org/GovOn.git
cd GovOn

# Backend (Python)
pip install -e ".[dev]"
pytest

# TUI (Node.js)
cd packages/npm && npm install && npm run build

See CONTRIBUTING.md for guidelines.

About

GovOn is an industry-academia project from the Department of Computer Engineering at Dong-A University. It is an AX (Agentic Transformation) platform that builds an agentic intelligence layer on top of Korea's existing DX (Digital Transformation) national infrastructure. Rather than replacing independent agency systems, GovOn wraps their APIs as LLM-callable tools — enabling a central agentic AI to orchestrate cross-agency workflows through a single interface. The AI does not replace human judgment -- it transforms fragmented digital infrastructure into a unified agentic experience so public servants can focus on decisions that matter.

License

The GovOn source code is licensed under MIT.

Important: The runtime uses EXAONE 4.0-32B-AWQ, which is distributed under the EXAONE AI Model License Agreement — non-commercial use only. Commercial deployment of the model requires a separate license from LG AI Research.