Skip to main content
Cherry Studio is useful for managing multiple model providers in one client. Add 4096bytes as an OpenAI compatible provider, then select its models in chats.

​
Prerequisites

  • You have completed Create an account.
  • You have completed Get an API Key.
  • Cherry Studio is installed and its settings page opens normally.
  • You have the 4096bytes API Key, Base URL, and model name.
Create a dedicated API Key for Cherry Studio, such as cherry-studio-home or cherry-studio-work.

​
Install

If Cherry Studio is not installed, open the Cherry Studio download page and choose the package for your system.
1

Download the package

Choose Windows, macOS, or Linux from the official download page. Windows users may need to choose between installer and portable builds. macOS users should choose Intel or Apple Silicon correctly.
2

Launch Cherry Studio

Open Cherry Studio once after installation and confirm the main window starts normally.
3

Open settings

Open Settings and find model service, provider, or API configuration pages.

​
Add a provider

1

Open settings

Launch Cherry Studio and open Settings. The entry point may be in the lower-left corner or sidebar.
2

Open model services

Find Model services, Model providers, Provider, or an API configuration page.
3

Add a provider

Click Add, New provider, or a similar action. Use 4096bytes as the provider name.
4

Choose compatible type

Choose OpenAI, OpenAI Compatible, Custom OpenAI, or a similar option.

​
Fill connection fields

Cherry Studio fieldValue
Name, Provider name4096bytes
API Key, KeyThe API Key you created in the 4096bytes dashboard.
API address, API Host, Base URLhttps://api.4096bytes.com.
API type, Compatible typeOpenAI compatible.
Model, Model nameThe model name shown in the dashboard.
Use the root Base URL only. Do not append /v1/chat/completions, /v1/responses, or another full endpoint path.
The API Key is stored in local Cherry Studio settings. Do not configure a personal key on shared computers, and do not share screenshots that expose the key.

​
Add models

Cherry Studio versions differ in how they handle model lists. Some can fetch models automatically; others require manual model names.
1

Try fetching models

If the page has Fetch models, Refresh model list, or Check, run it first.
2

Add a model manually

If the model list is empty, click Add model or Custom model and enter the model name shown in the dashboard.
3

Save provider

Enable the provider and save the API Key, Base URL, and model list.
The model name must match the dashboard exactly.

​
Test connection

After saving, create a new chat in Cherry Studio and select a model under the 4096bytes provider. Use a short test prompt: 
Tell me the model name you are currently using in one sentence.
If the model responds, Cherry Studio can send requests through 4096bytes.

​
Tips

  • Use a dedicated API Key for Cherry Studio.
  • If multiple providers are configured, check the selected model before sending a message.
  • If Cherry Studio supports model groups or default models, set your common 4096bytes model as the default.
  • Avoid personal keys on shared team devices. If sharing is necessary, use a separate key with a quota limit.

​
Troubleshooting

Confirm the provider type is OpenAI compatible, then check API Key, Base URL, and model name. Use https://api.4096bytes.com and do not fill a full endpoint path.
If Cherry Studio cannot fetch models automatically, add the model name from the dashboard manually. Save the provider and reopen the chat.
Check whether the API Key is complete and copied without extra spaces or newlines. Confirm the key was not deleted, expired, or over quota.
Copy the model name again from the dashboard. It must match exactly and must not come from another client page or old screenshot.
Check the current chat’s model selector. Cherry Studio may keep the last used model, so switch to the 4096bytes provider manually.
Start with a short prompt. Long context, attachments, proxy networks, and concurrent requests can affect response time. Check Cherry Studio logs if needed.