A highly efficient, secure, and stable application development platform with excellent performance, easy scalability, and deep integration of AI capabilities such as LLM, AI Chat, RAG, and Agents.

Language Options: 中文 | English

</div>

📋 Project Introduction

Weave (meaning "to weave") – from a simple thread to a complex tapestry, weaving is the creative process from simplicity to complexity. A high-performance, high-efficiency, easily extensible empowerment tool/AI application development platform built with Golang, designed for efficiently constructing stable and reliable intelligent applications. It employs a microkernel + layered architecture, allowing developers to develop efficiently and integrate/manage various tools/services with ease, while maintaining high system performance and scalability.

Weave's core strength lies in its stable and reliable AI feature development stack + plugin/tool system, deeply integrating core AI capabilities such as LLM (Large Language Models), AIChat (Intelligent Chat), Agent (Intelligent Agent), RAG (Retrieval-Augmented Generation), and more. It provides a unified development framework and standardized interfaces, enabling developers to rapidly build powerful intelligent applications without needing to focus on underlying complex implementations.

Primary application scenarios include:

Tool/Application Development Aggregation
Data/Service Flow Middle Platform
API Gateway and Service Orchestration
AI Application Development - Building LLM, AIChat, Agent, and RAG intelligent applications
Automated Workflows - Complex task automation combining ETL, DevOps, and Agent

🏗️ Overall Architecture

Weave adopts a microkernel + layered architectural design pattern, fully leveraging the advantages of both to ensure system availability, achieving high flexibility, scalability, and excellent performance.

Integration of Microkernel and Layered Architecture

Weave incorporates the design philosophy of layered architecture on top of a microkernel architecture, forming a complete, efficient, and flexible architectural system:

Microkernel Architecture (Plugin System): Provides plugin management, lifecycle control, and inter-plugin communication mechanisms.
Layered Architecture (Core System): Separates core functionalities by concerns, forming a clear hierarchical structure.

Microkernel Architecture Components

Core Kernel: Provides fundamental runtime environment, plugin management, configuration management, logging services, security mechanisms, and other basic functionalities.
Plugin System: The plugin manager handles plugin registration, lifecycle management, dependency resolution, and conflict detection.
Extension Plugins: Integrated into the core system through plugin interfaces, implementing various business functionalities.

Layered Architecture Components

Interface Layer: Handles HTTP requests, including routing management and controllers.
Business Layer: Contains core business logic and the plugin system.
Data Layer: Responsible for data storage and access.
Infrastructure Layer: Provides services such as logging, configuration, and security.

Architectural Characteristics

Loosely Coupled Design: The core system and plugins communicate through well-defined interfaces, reducing inter-module dependencies.
Hot-Plug Capability: Plugins can be dynamically loaded and unloaded at runtime without system restart.
Feature Isolation: Each plugin independently encapsulates functionality, with its own namespace and route prefix.
Dependency & Conflict Management: Built-in dependency resolution and conflict detection mechanisms ensure harmonious coexistence among plugins.
Unified Interface: All plugins implement the same Plugin interface, standardizing the development process.
Extensibility: System functionalities can be extended on-demand without modifying kernel code.
Clear Hierarchy: The core system uses layered design for rational code organization, making it easy to maintain and extend.
High Performance: The layered design optimizes request processing flow, improving system response speed.

The system's core is an efficient and stable plugin mechanism and service aggregation, allowing functional modules to be independently developed and deployed as plugins/services, while interacting through unified interfaces. The overall architectural design emphasizes modularity, scalability, and high performance.

🌟 Project Features

🏗️ Microkernel + Layered Architecture

Stable Core & Clear Hierarchy: The core system remains minimal, and the layered design ensures rational code organization, making it easy to maintain and extend.
Flexible Functional Extension: Extend system functionalities on-demand through the plugin mechanism without modifying kernel code.
Low Coupling, High Cohesion: System components are loosely coupled, facilitating maintenance and upgrades.
Hot-Plug Capability: Plugins can be dynamically loaded and unloaded at runtime without restarting the system.
Feature Isolation & Unified Management: Each plugin independently encapsulates functionality, with its own namespace and route prefix, while core services are uniformly managed through the layered architecture.
Dependency & Conflict Management: Built-in dependency resolution and conflict detection mechanisms ensure harmonious plugin coexistence.
Unified Interface: All plugins implement the same Plugin interface, standardizing the development process.
High Performance: The layered design optimizes the request processing flow, improving system response speed.

🚀 High Performance/Efficiency

Built on the Gin framework, offering high performance and strong concurrency capabilities.
Optimized database connection pool supporting high-concurrency access.
Modular architectural design with a clear code structure, making it easy to maintain and extend.
Supports environment variable overrides for easy configuration across different environments.
Efficient routing management supporting dynamic routes and parameter binding.
The layered architecture optimizes request processing flow, enhancing system response speed.

🔌 Plug-in Easy Extensibility

Unified plugin interface design supporting hot-plugging.
Plugin manager uniformly registers, manages, and executes plugins.
Plugins can independently register routes with their own namespace.
Plugin dependency and conflict detection mechanisms.
Scaffolding tool for conveniently generating plugin framework code.
Example plugins demonstrating the complete plugin development process.

🧠 Deep AI Integration

Weave provides a complete AI development functional stack, deeply integrating core AI technologies such as LLM, AIChat, Agent, and RAG, offering developers a one-stop intelligent application development experience.

Complete AI Functional Stack: Integrates core AI capabilities like LLM (Large Language Models), AIChat (Intelligent Chat), Agent (Intelligent Agent), and RAG (Retrieval-Augmented Generation), covering the entire process from basic model invocation to complex intelligent application development.
Multi-Model Support: Compatible with mainstream LLM platforms like OpenAI, Ollama, ModelScope, etc., supporting dynamic switching of models and platforms to meet different scenario needs.
Comprehensive AIChat Services: Multi-platform, multi-model, multimodal support, streaming response, context optimization, summary extraction, conversation management, and other features.
Intelligent Agent Framework: Provides a complete Agent development framework supporting tool calling, task planning, complex workflow automation, and multi-turn conversation management.
Efficient RAG System: Implements high-performance vector search based on RedisSearch, supports multi-format document processing (text, PDF, Markdown, etc.), improving the accuracy and relevance of generated content.
Flexible Embedding Models: Supports custom embedding models and retrieval parameters to adapt to different business scenarios' vector representation needs.
Seamless Architectural Integration: AI capabilities can be used as service aggregates or plugins, facilitating rapid construction and extension of intelligent applications.

🔒 Secure & Reliable

JWT-based authentication and authorization system.
Comprehensive CSRF protection mechanism.
Rate-limiting middleware based on the token bucket algorithm.
Password hashing storage and verification.
Detailed login history records.
Unified error handling middleware.
HTTPS support (can be enabled in configuration).
The layered architecture encapsulates security mechanisms uniformly in the infrastructure layer for easy unified management and maintenance.

📊 Observability

Integrated structured logging system (z

Weave

Install / Use

README