Repository: coze-dev/coze-studio

Stars: 20548

CLAUDE.md Copy Chapter

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project Overview Copy Chapter

Coze Studio is an all-in-one AI agent development platform with both frontend (React + TypeScript) and backend (Go) components. The project uses a sophisticated monorepo architecture managed by Rush.js with 135+ frontend packages organized in a hierarchical dependency system.

Development Commands Copy Chapter

Environment Setup Copy Section

Clone and setup

git clone https://github.com/coze-dev/coze-studio.git
cd coze-studio
Install frontend dependencies

rush update
For Docker-based development

cd docker
cp .env.example .env
Configure model settings in backend/conf/model/

docker compose up -d
Access at http://localhost:8888

Development Workflow Copy Section

Start middleware services (MySQL, Redis, Elasticsearch, etc.)

make middleware
Start Go backend in development mode

make server
Start frontend development server

cd frontend/apps/coze-studio
npm run dev
Full development environment

make debug

Build Commands Copy Section

Build frontend only

make fe
Build Go server

make build_server
Build everything with Docker

make web
Rush monorepo commands

rush build                    # Build all packages
rush rebuild -o @coze-studio/app  # Build specific package
rush test                     # Run all tests
rush lint                     # Lint all packages

Testing Copy Section

Run tests (Vitest-based)

rush test
npm run test                  # In specific package
npm run test:cov             # With coverage
Backend tests

cd backend && go test ./...

Architecture Overview Copy Chapter

Frontend Architecture Copy Section

Backend Architecture (Go) Copy Section

Key Architectural Patterns Copy Section

Database & Infrastructure Copy Chapter

Docker Services Stack Copy Section

Database Management Copy Section

Sync database schema

make sync_db
Dump database schema

make dump_db
Initialize SQL data

make sql_init
Atlas migration management

make atlas-hash

Key Development Patterns Copy Chapter

Frontend Package Development Copy Section

Backend Development Copy Section

Model Configuration Copy Section

Testing Strategy Copy Chapter

Coverage Requirements by Package Level Copy Section

Testing Framework Copy Section

Common Issues & Solutions Copy Chapter

Frontend Development Copy Section

Backend Development Copy Section

Docker Issues Copy Section

IDL and Code Generation Copy Chapter

The project uses Interface Definition Language (IDL) for API contract management:
- IDL files in idl/ directory (Thrift format)
- Frontend code generation via @coze-arch/idl2ts-* packages
- Backend uses generated Go structs

Plugin Development Copy Chapter

For custom plugin development:
- Reference templates in backend/conf/plugin/pluginproduct/
- Follow OAuth schema in backend/conf/plugin/common/oauth_schema.json
- Configure authentication keys for third-party services

Contributing Copy Chapter

- Use conventional commits via rush commit
- Run linting with rush lint-staged (pre-commit hook)
- Ensure tests pass before submitting PRs
- Follow team-based package organization and tagging conventions

README.md Copy Chapter

!Image

<div align="center"><p>
<a href="#what-is-coze-studio">Coze Studio</a> •
<a href="#feature-list">Feature list</a> •
<a href="#quickstart">Quickstart</a> •
<a href="#developer-guide">Developer Guide</a>
</p>
<p>
<img alt="License" src="https://img.shields.io/badge/license-apache2.0-blue.svg">
<img alt="Go Version" src="https://img.shields.io/badge/go-%3E%3D%201.23.4-blue">
</p>

English | 中文

</div>

What is Coze Studio? Copy Chapter

Coze Studio is an all-in-one AI agent development tool. Providing the latest large models and tools, various development modes and frameworks, Coze Studio offers the most convenient AI agent development environment, from development to deployment.

* Provides all core technologies needed for AI agent development: prompt, RAG, plugin, workflow, enabling developers to focus on creating the core value of AI.
* Ready to use for professional AI agent development at the lowest cost: Coze Studio provides developers with complete app templates and build frameworks, allowing you to quickly construct various AI agents and turn creative ideas into reality.

Coze Studio, derived from the "Coze Development Platform" which has served tens of thousands of enterprises and millions of developers, we have made its core engine completely open. It is a one-stop visual development tool for AI Agents that makes creating, debugging, and deploying AI Agents unprecedentedly simple. Through Coze Studio's visual design and build tools, developers can quickly create and debug agents, apps, and workflows using no-code or low-code approaches, enabling powerful AI app development and more customized business logic. It's an ideal choice for building low-code AI products tailored . Coze Studio aims to lower the threshold for AI agent development and application, encouraging community co-construction and sharing for deeper exploration and practice in the AI field.

The backend of Coze Studio is developed using Golang, the frontend uses React + TypeScript, and the overall architecture is based on microservices and built following domain-driven design (DDD) principles. Provide developers with a high-performance, highly scalable, and easy-to-customize underlying framework to help them address complex business needs.

Feature list Copy Chapter

Quickstart Copy Chapter

Environment requirements:

* Before installing Coze Studio, please ensure that your machine meets the following minimum system requirements: 2 Core、4 GB
* Pre-install Docker and Docker Compose, and start the Docker service.

Deployment steps:

1. Retrieve the source code.

# Clone code
   git clone https://github.com/coze-dev/coze-studio.git

cd coze-studio
   # start service
   # for macOS or Linux
   make web  
   # for windows
   cp ./docker/.env.example ./docker/.env
   docker compose -f ./docker/docker-compose.yml up

For common startup failure issues, please refer to the FAQ.
3. Register an account by visiting http://localhost:8888/sign, entering your username and password, and clicking the Register button.
4. Configure the model at http://localhost:8888/admin/#model-management by adding a new model. (The image version must be greater than or equal to 0.5.0.)
5. Visit Coze Studio at http://localhost:8888/.

Developer Guide Copy Chapter

* Project Configuration:
* Model Configuration: Before deploying the open-source version of Coze Studio, you must configure the model service. Otherwise, you cannot select models when building agents, workflows, and apps.
* Plugin Configuration: To use official plugins from the plugin store, you must first configure the plugins and add the authentication keys for third-party services.
* Basic Component Configuration: Learn how to configure components such as image uploaders to use functions like image uploading in Coze Studio .
* API Reference: The Coze Studio Community Edition API and Chat SDK are authenticated using Personal Access Token, providing APIs for conversations and workflows.
* Development Guidelines:
* Project Architecture: Learn about the technical architecture and core components of the open-source version of Coze Studio.
* Code Development and Testing: Learn how to perform secondary development and testing based on the open-source version of Coze Studio.
* Troubleshooting: Learn how to view container states and system logs.

Using the open-source version of Coze Studio Copy Chapter

Regarding how to use Coze Studio, refer to the Coze Development Platform Official Documentation Center for more information. Please note that certain features, such as tone customization, are limited to the commercial version. Differences between the open-source and commercial versions can be found in the Feature List.

* Quick Start: Quickly build an AI assistant agent with Coze Studio.
* Developing Agents: Learn how to create, build, publish, and manage agents. You can use functions such as knowledge, plugins, etc., to resolve model hallucination and lack of expertise in professional fields. In addition, Coze Studio provides rich memory features that enable agents to generate more accurate responses based on a personal user's historical conversations during interactions.
* Develop workflows: A workflow is a set of executable instructions used to implement business logic or complete specific tasks. It structures data flow and task processing for apps or agents. Coze Studio provides a visual canvas where you can quickly build workflows by dragging and dropping nodes.
* Resources such as plugins: In Coze Studio, workflows, plugins, databases, knowledge bases, and variables are collectively referred to as resources.
* API & SDK: Coze Studio supports API related to chat and workflows, and you can also integrate agents or apps with local business systems through Chat SDK.
* Tutorials for practice: Learn how to use Coze Studio to implement various AI scenarios, such as building web-based online customer service using Chat SDK.

License Copy Chapter

Community contributions Copy Chapter

Security and privacy Copy Chapter

Join Community Copy Chapter

We are committed to building an open and friendly developer community. All developers interested in AI Agent development are welcome to join us!

🐛 Issue Reports & Feature Requests Copy Section

💬 Technical Discussion & Communication Copy Section

Feishu Group Chat
Scan the QR code below with Feishu mobile app to join:

!Image

Discord Server
Click to join: Coze Community

Telegram Group
Click to join: Telegram Group Coze

Acknowledgments Copy Chapter

* The Eino framework team - providing powerful support for Coze Studio's agent and workflow runtime engines, model abstractions and implementations, and knowledge base indexing and retrieval
* The FlowGram team - providing a high-quality workflow building engine for Coze Studio's frontend workflow canvas editor
* The Hertz team - Go HTTP framework with high-performance and strong-extensibility for building micro-services
* All users who participated in testing and feedback