Product Overview

NiteMoon AI Platform is an enterprise-grade AI application platform built entirely in Java. It integrates a visual workflow engine, multi-model LLM, RAG knowledge base, knowledge graph, YOLO vision platform, digital human / TTS, and other core capabilities, providing enterprises and developers with a one-stop AI application development solution.

Core Capabilities Matrix

Visual Workflow Engine

Drag-and-drop AI workflow construction, supporting 13 node types for complex AI pipelines. Supports conditional branching, context passing, token statistics, and more.

• 13 built-in node types
• Visual flow orchestration
• Conditional branching and loops
• SSE streaming response

Learn More →

Multi-Model LLM Integration

Unified interface to connect 15+ mainstream large language models, supporting text chat, embedding generation, image generation, web search, and other model types.

• 15+ LLM providers
• 4 model types
• Strategy pattern abstraction
• Dynamic model switching

Learn More →

RAG Knowledge Base

Dual-mode retrieval with vector search and graph search for precise knowledge context retrieval. Supports document upload, parsing, chunking, and embedding generation.

• 3 vector store backends
• Intelligent document parsing
• Configurable retrieval parameters
• Streaming conversation support

Learn More →

Knowledge Graph

Neo4j-powered intelligent knowledge network with LLM-based automatic entity-relation extraction, graph traversal retrieval, and 1-hop neighbor expansion.

• LLM entity-relation extraction
• HanLP Chinese tokenization
• Graph traversal retrieval
• Entity deduplication and confidence scoring

Learn More →

YOLO Vision Platform

The first complete YOLO lifecycle management platform built in Java, supporting training, inference, and real-time video stream detection.

• YOLOv8 model training & monitoring
• ONNX model inference
• RTMP/RTSP/HTTP video stream detection
• Dynamic model upload
• Car plate detection

Learn More →

Digital Human / TTS

Multi-engine TTS support with WebSocket bidirectional communication for an intelligent voice interaction experience.

• 3 TTS engines
• WebSocket bidirectional protocol
• Voice file generation
• Digital human configuration management

Learn More →

AI Excel Generation

Natural language-driven intelligent report generation with a conversational Excel creation interface for improved data analysis efficiency.

• LLM-powered generation
• Conversational interface
• Template management
• Multi-format export

Learn More →

Deployment Modes Comparison

NiteMoon AI Platform supports two deployment modes, flexibly adapting to different scale requirements:

Feature	Monolith Mode (boot)	Microservices Mode (cloud)
Use Cases	Development, testing, small-scale deployment	Production, large-scale deployment
Deployment Complexity	Low	High
Scalability	Limited	High
Resource Usage	Low	High
Dependencies	MySQL, Redis	MySQL, Redis, Nacos, RocketMQ
Service Communication	Local calls	Dubbo RPC
Configuration	Local configuration	Nacos configuration center

Recommendation

• Monolith Mode: Ideal for individual developers, small projects, and rapid prototyping
• Microservices Mode: Ideal for enterprise applications, high-concurrency scenarios, and horizontal scaling needs

Technical Architecture

View Technical Architecture Details

Integration Capabilities

NiteMoon AI Platform offers rich integration capabilities:

File Storage

• MinIO (local / private cloud)
• Alibaba Cloud OSS
• Qiniu Cloud
• Tencent Cloud COS
• AWS S3
• FTP/SFTP

Message Queue

• RocketMQ

Monitoring

• Spring Boot Admin

API Documentation

• Knife4j / SpringDoc OpenAPI

Authentication & Authorization

• Sa-Token RBAC
• WeChat Mini Program Login