Expand description
MoFA API - Standard SDK for MoFA framework
This crate provides a standardized API for the MoFA (Model-based Framework for Agents) framework.
§Architecture Layers
The SDK is organized into clear layers following microkernel architecture principles:
┌─────────────────────────────────────────┐
│ User Code │
└─────────────────┬───────────────────────┘
↓
┌─────────────────────────────────────────┐
│ SDK (Standard API Surface) │
│ - kernel: Core abstractions │
│ - runtime: Lifecycle management │
│ - foundation: Business functionality │
└─────────────────┬───────────────────────┘§Features
dora- Enable dora-rs runtime support for distributed dataflow
For FFI bindings (Python, Kotlin, Swift, Java), use the mofa-ffi crate.
§Quick Start
mofa-sdk = "0.1"ⓘ
use mofa_sdk::kernel::{AgentInput, MoFAAgent};
use mofa_sdk::runtime::run_agents;
struct MyAgent;
#[async_trait::async_trait]
impl MoFAAgent for MyAgent {
// implementation...
}
#[tokio::main]
async fn main() -> anyhow::Result<()> {
let outputs = run_agents(MyAgent, vec![AgentInput::text("Hello")]).await?;
println!("{}", outputs[0].to_text());
Ok(())
}Modules§
- agent
- Agent building blocks and concrete implementations (foundation layer)
- collaboration
- 自适应协作协议模块
- config
- Global configuration facade (kernel + runtime + foundation)
- coordination
- Coordination strategies and schedulers (foundation layer)
- foundation
- Business functionality and concrete implementations
- kernel
- Core agent abstractions and extensions
- llm
- LLM (Large Language Model) integration module
- messaging
- Generic message bus framework for decoupled agent architectures
- persistence
- plugins
- prelude
- Commonly used types for quick start
- prompt
- Prompt templates, registries, and composition utilities
- react
- ReAct (Reasoning + Acting) 框架
- rhai
- runtime
- Agent lifecycle and execution management
- secretary
- 秘书Agent模式 - 基于事件循环的智能助手
- skills
- Agent Skills 管理 API
- workflow
- Workflow orchestration module with LangGraph-inspired StateGraph API
Functions§
- anthropic_
from_ env - 从环境变量创建 Anthropic 提供器
- gemini_
from_ env - 从环境变量创建 Google Gemini 提供器