Skip to content

2. Quickstart

Jump to bottom
newacc edited this page Aug 11, 2025 · 13 revisions

Requirements

Before installing Coze Studio, ensure that your software and hardware environment meets the following requirements:

Requirements Description
CPU 2 Core
RAM 4 GiB
Docker Install Docker and Docker Compose in advance, and start the Docker service. For detailed instructions, refer to the Docker documentation:

* macOS: It is recommended to use Docker Desktop for installation. Refer to the Docker Desktop For Mac installation guide.
* Linux: Refer to the Docker installation guide and the Docker Compose installation guide.
* Windows: It is recommended to install using Docker Desktop. Refer to the Docker Desktop For Windows installation guide.

Install Coze Studio

Step 1: Clone the source code

Run the following commands in your local project to clone the latest version of the Coze Studio source code.

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

Step 2: Configure a model

Coze Studio is an AI app development platform based on large language models (LLMs). Before deploying and starting the open-source version of Coze Studio for the first time, you need to configure the model service in your Coze Studio project; otherwise, you will not be able to properly select a model when creating an agent or workflow. This document uses the Volcengine Ark model as an example to demonstrate how to configure a model service for Coze Studio. If you plan to use OpenAI or other online model services, you should refer to the model configuration documentation to correctly fill in the configuration file.

  1. Copy the template files for the doubao-seed-1.6 model from the template directory and paste them into the configuration file directory.

    cd coze-studio
Copy a model configuration template
cp backend/conf/model/template/model_template_ark_doubao-seed-1.6.yaml backend/conf/model/ark_doubao-seed-1.6.yaml

  2. In the configuration file directory, modify the template file.

    1. Go to the backend/conf/model directory. Open the fileark_doubao-seed-1.6.yaml.
    2. Set the id, meta.conn_config.api_key, and meta.conn_config.model fields, and save the file. Image

For users in China, you may use Volcengine Ark; for users outside China, you may use BytePlus ModelArk instead.

Step 3: Deploy and start services

The initial deployment and startup of Coze Studio require retrieving images and building local images. This process may take some time, so please be patient. During the deployment process, you will see the following log information. If you see the message "Container coze-server Started", it means that the Coze Studio service has started successfully.

Start the service
cd docker
cp .env.example .env
docker compose --profile '*' up -d

After the service starts, the coze-elasticsearch-setup, coze-minio-setup, coze-mysql-setup-init-sql, and coze-mysql-setup-schema containers are in an exited state (exit 0), which is normal.

Step 4: Log in

After starting the service, use a browser to visit http://localhost:8888/ to open Coze Studio. 8888 is the backend listening port.

  1. Enter your email and password.
  2. Click the Register button to complete registration. After registration, the page will automatically log you in, and you can start experiencing all the features and services of Coze Studio. At this point, you have successfully deployed and logged in to Coze Studio.

Image

What's next

After successfully deploying Coze Studio, if you need to use functions such as plugins and knowledge bases, you also need to:

  • Configure Plugins: Some official plugins require authentication through the keys of third-party services, such as Feishu Cloud Document series components. If the key is not configured, the plugin will be displayed as "Unauthorized".
  • Configure Basic Components: The core components are as follows
    • Knowledge Base: To use the knowledge base function, you must configure the Embedding Component; for image knowledge bases, you also need to set up the OCR Component to recognize text in images.
    • Image Upload: When you need to use the multimodal input of the large model, the upload component needs to be configured with a public network domain name or IP address. Otherwise, in the debug console and when conversing with the model, the model cannot read the uploaded images.
  • Configure Models: Add model services as needed so that your agents, workflows or applications can use more models.

Security risks in public networks

If Coze Studio is to be deployed in a public network environment, it is recommended to pay attention to the following security risks:

  • It is recommended to turn off the registration function or enable the email whitelist; otherwise, any user can use it via the link.
  • It is recommended to enable the default sandbox environment for the workflow code nodes to enhance security. For detailed configuration instructions, please refer to the Sandbox Environment for Code Nodes.
  • It is recommended to configure the network for the deployment environment according to business requirements (such as intranet access) to avoid SSRF risks.
  • By default, the Coze Studio server only listens to localhost. When deployed in a public network environment, it is recommended to listen to 0.0.0.0 only when necessary or add additional security measures to prevent the service from being directly exposed to the public network.

Home

Clone this wiki locally