# Connect Cherry Studio to NezhaGate: 10-Minute Setup

A beginner-proof walkthrough: add NezhaGate as a provider in Cherry Studio, pick models, verify the connection and chat. Model picks and error fixes included.

> https://nezhagate.com/en/docs/integrations/cherry-studio

## What you need

- Cherry Studio 2.x for desktop (Windows, macOS or Linux), downloaded from the Cherry Studio website or GitHub.

- A NezhaGate account and API key (step 1 shows you how).

- About 10 minutes.

## Step 1: Sign up for NezhaGate and create an API key

1. Open [nezhagate.com](https://nezhagate.com/en/) and click **Sign up free** in the top-right corner. Email works, and so does one-click sign-in with Google or GitHub.

2. Go to **Console → API Keys**. Under **Create API Key**, name the key (for example `cherry`), leave **Per-Key Limit** at 0 for no limit, and click **Create Key**.

3. A key starting with `sk-` appears. **Copy it right away**: it is shown only once.

Check and top up your balance on the **Billing** page. You pay only for what you use, and failed requests are never billed. Per-model rates are on the [pricing page](https://nezhagate.com/en/pricing).

## Step 2: Add NezhaGate as a provider

1. Open Cherry Studio, click the gear icon (**Settings**) at the bottom of the left bar and go to **Model Provider**.

2. Below the provider list, click **+ Add Provider**.

3. In the **Add Custom Provider** dialog, fill in:

   - **Provider Name**: `NezhaGate`

   - **API Key**: the key you copied in step 1

   - Under **Endpoint settings**, the **OpenAI** field: the address below. Leave the **Anthropic** field empty.

```
https://nezhagate.com/v1
```

1. Under the field you will see "Request path: https://nezhagate.com/v1/chat/completions". If it matches, click **Add**.

## Step 3: Pick models, verify, enable

1. The **Choose models** dialog opens next, filled from NezhaGate's model list.

2. **Don't click Select All**: the list also has image and video models, which do not work in chat. Tick just the chat models you want, for example `deepseek-v4.1-flash`, `kimi-k3`, `glm-5.3`, `qwen3.7-max` and `claude-opus-5`.

3. Click **Add selected models**. Cherry Studio verifies a model (one real request), enables the provider, and you finish with **Done**.

Don't click **Skip**. Skipping leaves the provider switched off, and you would have to turn on the switch at the top right of the provider page yourself. An enabled provider shows a green dot in the list.

To add models later, open the NezhaGate provider page and click **Sync models**; for a model that is not listed, use **Add model manually** and type its ID. To test the connection again, click **Model Check**, pick a model and press **Start**: you should see "Connection successful".

## Step 4: Start chatting

1. Go back to the chat screen, open the model picker and choose a model under **NezhaGate**.

2. Send "Hello". Once a reply comes back, you are all set.

## Recommended models

| Use | Model ID | Notes |
| --- | --- | --- |
| Everyday chat, translation, summaries | `deepseek-v4.1-flash` | Cheap and fast; the best value. |
| Long documents, many files | `kimi-k3` | Strong long-context understanding; prompt caching makes repeat questions about the same document cheaper. |
| Coding and hard reasoning | `glm-5.3`, `qwen3.7-max` | Both think before answering: a little slower, more accurate. |
| The hardest tasks | `claude-opus-5`, `gpt-6-astra` | Flagship models, priced higher. |
| Fiction and long-form writing | `claude-fable-5` | The strongest at narrative and long-form prose. |

Current prices are on the [pricing page](https://nezhagate.com/en/pricing).

## Getting the address right

- Both `https://nezhagate.com/v1` and `https://nezhagate.com` work: without a version in the path, Cherry Studio adds `/v1` itself.

- Do **not** enter `/v1/chat/completions`. The path would be doubled and you would get a 404.

- A `#` at the end of the address tells Cherry Studio not to add a version. You do not need it for NezhaGate.

Still on 1.x? The screens differ: when adding the provider, set **Provider Type** to OpenAI and **API Host** to `https://nezhagate.com`, then use **Manage** to add models and **Check** to test. Upgrading to 2.x is the easier path.

## Common problems and errors

| What you see | What to do |
| --- | --- |
| "The provider returned no available models" or "Failed to pull models" | Check the address and key (no spaces around the key), or use **Add model manually** and type the model ID. |
| "The model request failed" during verification | Make sure your account has a balance. Thinking models such as `qwen3.7-max` can be slow enough to time out the check; verify with `deepseek-v4.1-flash` instead. |
| NezhaGate models are missing from the chat model picker | The provider is off. Go to Settings → Model Provider → NezhaGate and turn on the switch at the top right. |
| 404 | The address has `/v1` twice or ends in `/chat/completions`. Change it to `https://nezhagate.com/v1`. |
| 401 | The key is wrong or was deleted. Create a new one in the console and copy it. |
| 402 | Your balance ran out, or the key hit the limit you set. Top up on the Billing page or raise the key's limit. |
| 400 | Usually an unsupported parameter, too much context or an oversized image. Try again in a new assistant with default settings. Sending an image to a model without vision also returns 400. |

Still stuck? See the [API docs](https://nezhagate.com/en/docs-guide) or the [FAQ](https://nezhagate.com/en/faq).