Skip to content

Commit 10268df

Browse files
committed
docs: update README.md
1 parent 00f8008 commit 10268df

File tree

2 files changed

+26
-26
lines changed

2 files changed

+26
-26
lines changed

README.md

Lines changed: 13 additions & 13 deletions
Original file line numberDiff line numberDiff line change
@@ -6,7 +6,7 @@ AgentMesh is a **Multi-Agent platform** for AI agents building, providing a fram
66
task planning, and autonomous decision-making. Build your agent team quickly and solve complex tasks through agent
77
collaboration.
88

9-
# Overview
9+
## Overview
1010

1111
AgentMesh uses a modular layered design for flexible and extensible multi-agent systems:
1212

@@ -18,15 +18,15 @@ AgentMesh uses a modular layered design for flexible and extensible multi-agent
1818
protocol support coming soon for even more tool extensions.
1919
- **Multi-Platform**: Run via CLI, Docker, or SDK. WebUI and integration with common software coming soon.
2020

21-
# Quick Start
21+
## Quick Start
2222

2323
Choose one of these three ways to build and run your agent team:
2424

25-
## 1. Terminal
25+
### 1. Terminal
2626

2727
Run a multi-agent team from your command line:
2828

29-
### 1.1 Installation
29+
#### 1.1 Installation
3030

3131
**Requirements:** Linux, MacOS, or Windows with Python installed.
3232

@@ -53,7 +53,7 @@ pip install browser-use
5353
playwright install
5454
```
5555

56-
### 1.2 Configuration
56+
#### 1.2 Configuration
5757

5858
Edit the `config.yaml` file with your model settings and agent configurations:
5959

@@ -66,7 +66,7 @@ Add your model `api_key` - AgentMesh supports `openai`, `claude`, `deepseek`, `q
6666
> The template includes a pre-configured `software_team` with three roles (PM, architect, engineer) that collaborate on
6767
> software development tasks.
6868
69-
### 1.3 Execution
69+
#### 1.3 Execution
7070

7171
```bash
7272
python main.py -l # List available agent teams
@@ -75,7 +75,7 @@ python main.py -t software_team # Run the 'software_team'
7575

7676
Enter your requirements in the interactive prompt to begin.
7777

78-
## 2. SDK
78+
### 2. SDK
7979

8080
Use the AgentMesh SDK to build custom agent teams programmatically:
8181

@@ -106,13 +106,13 @@ team.add(Agent(name="Developer", description="Implements code based on requireme
106106
result = team.run(task="Write a Snake client game")
107107
```
108108

109-
## 3. Web Service
109+
### 3. Web Service
110110

111111
Coming soon
112112

113-
# Details
113+
## Details
114114

115-
## Core Concepts
115+
### Core Concepts
116116

117117
- **Agent**: Autonomous decision-making unit with specific roles and capabilities, configurable with models, system
118118
prompts, tools, and decision logic.
@@ -122,22 +122,22 @@ Coming soon
122122
- **Context**: Shared information including team details, task content, and execution history.
123123
- **LLMModel**: Large language model interface supporting various mainstream LLMs through a unified API.
124124

125-
## Supported Models
125+
### Supported Models
126126

127127
- **OpenAI**: GPT series models, recommended: `gpt-4.1`, `gpt-4o`, `gpt-4.1-mini`
128128
- **Claude**: Claude series models, recommended: `claude-3-7-sonnet-latest`
129129
- **DeepSeek**: DeepSeek series models, recommended: `deepseek-chat`
130130
- **Ollama**: Local open-source models (coming soon)
131131

132-
## Built-in Tools
132+
### Built-in Tools
133133

134134
- **calculator**: Mathematical calculation tool supporting complex expression evaluation
135135
- **current_time**: Current time retrieval tool solving model time awareness issues
136136
- **browser**: Web browsing tool based on browser-use, supporting web access, content extraction, and interaction
137137
- **google_search**: Search engine tool for retrieving up-to-date information
138138
- **MCP**: Multi-modal content processing tool supporting image recognition and generation (coming soon)
139139

140-
# Contribution
140+
## Contribution
141141

142142
Star this project to receive notifications about updates.
143143

docs/README-CN.md

Lines changed: 13 additions & 13 deletions
Original file line numberDiff line numberDiff line change
@@ -6,7 +6,7 @@
66
AgentMesh是一个开源的 **多智能体 (Multi-agent) 平台** ,提供开箱即用的Agent开发框架、多Agent间的协同策略、任务规划和自主决策能力。
77
在该平台上可以快速构建你的Agent团队,通过多Agent之间的协同完成任务。
88

9-
# 概述
9+
## 概述
1010

1111
AgentMesh 采用模块化分层设计,提供灵活且可扩展的多智能体系统构建能力:
1212

@@ -15,15 +15,15 @@ AgentMesh 采用模块化分层设计,提供灵活且可扩展的多智能体
1515
- 可扩展工具:内置搜索引擎、浏览器、文件系统、终端等工具,并将通过支持 MCP 协议获得更多工具扩展
1616
- 多端运行:支持命令行、Docker、SDK 等多种运行方式,即将支持 WebUI 及多种常用软件的集成
1717

18-
# 快速开始
18+
## 快速开始
1919

2020
提供三种使用方式快速构建并运行你的 Agent Team:
2121

22-
## 一、终端运行
22+
### 1.终端运行
2323

2424
在终端中命令行中快速运行多智能体团队:
2525

26-
### 1.安装
26+
#### 1.1 安装
2727

2828
**环境准备:** 支持 Linux、MacOS、Windows 系统,需要安装 python。
2929

@@ -50,7 +50,7 @@ pip install browser-use
5050
playwright install
5151
```
5252

53-
### 2.配置
53+
#### 1.2 配置
5454

5555
配置文件为根目录下的 `config.yaml`,包含模型配置和Agent配置,可以从模板文件复制后修改:
5656

@@ -62,7 +62,7 @@ cp config-template.yaml config.yaml
6262

6363
配置模板中预置了一个名为 `software_team` 的Agent开发团队,包含产品经理、架构师、工程师三种角色,可以协作完成软件开发任务。
6464

65-
### 3.运行
65+
#### 1.3 运行
6666

6767
```bash
6868
python main.py -l # 查看可用agent team
@@ -71,7 +71,7 @@ python main.py -t software_team # 运行名为 'software_team' 的team
7171

7272
进入交互模式后输入需求内容即可开始运行。
7373

74-
## 二、 SDK开发
74+
### 二、 SDK开发
7575

7676
`Agentmesh`的核心模块通过SDK对外提供,开发者可基于该SDK快速构建智能体及多智能体团队。
7777

@@ -104,13 +104,13 @@ team.add(Agent(name="Developer", description="Implements code based on PRD and a
104104
team.run(task="Write a Snake client game")
105105
```
106106

107-
## 三、Web服务运行
107+
### 三、Web服务运行
108108

109109
即将支持
110110

111-
# 详细介绍
111+
## 详细介绍
112112

113-
## 概念
113+
### 概念
114114

115115
- Agent: 智能体,具有特定角色和能力的自主决策单元,可配置模型、系统提示词、工具集和决策逻辑
116116
- AgentTeam: 智能体团队,由多个Agent组成,负责任务分配、上下文管理和协作流程控制
@@ -119,22 +119,22 @@ team.run(task="Write a Snake client game")
119119
- Context: 上下文,包含团队信息、任务内容和Agent间共享的执行历史
120120
- LLMModel: 大语言模型,支持多种主流大语言模型,统一接口设计
121121

122-
## 模型
122+
### 模型
123123

124124
- OpenAI: 支持 GPT 系列模型,推荐使用 `gpt-4.1`, `gpt-4o`, `gpt-4.1-mini`
125125
- Claude: 支持 Claude系列模型,推荐使用 `claude-3-7-sonnect-latest`
126126
- DeepSeek: 支持 DeepSeek 系列模型,推荐使用 `deepseek-chat`
127127
- Ollama: 支持本地部署的开源模型 (即将支持)
128128

129-
## 工具
129+
### 工具
130130

131131
- calculator: 数学计算工具,支持复杂表达式求值
132132
- current_time: 获取当前时间工具,解决模型时间感知问题
133133
- browser: 浏览器操作工具,基于browser-use实现,支持网页访问、内容提取和交互操作
134134
- google_search: 搜索引擎工具,获取最新信息和知识
135135
- MCP: 多模态内容处理工具,支持图像识别和生成(即将支持)
136136

137-
# 贡献
137+
## 贡献
138138

139139
Star支持和关注本项目,可以接受最新的项目更新通知。
140140

0 commit comments

Comments
 (0)