# DeepChat + QCode on Windows 11: step-by-step setup


**Short answer**: In DeepChat’s **Settings → Providers → Provider Center**, click **+ Add Custom Provider**, set **Name** to `QCode`, **API Type** to `OpenAI Completions`, **API Base URL** to `https://api.qcode.cc/openai/v1`, paste your `cr_` API key, and click **Connect and load models** (for Claude, add a second custom provider with **API Type** `Anthropic` and **API Base URL** `https://api.qcode.cc/api`).

| Item | Fill in |
|---|---|
| Provider name | `QCode` |
| API Type (GPT) | `OpenAI Completions` |
| API Base URL (GPT) | `https://api.qcode.cc/openai/v1` |
| API Base URL (Claude) | `https://api.qcode.cc/api` |
| API Key | `cr_xxxxxxxx` |

Checked on 2026-11-22 · DeepChat 1.1.4 · Windows 11


Follow the 11 screenshots below to install DeepChat on Windows 11, add QCode as a custom provider, and chat with `gpt-6.2-sol` (or Claude models) without creating any vendor account.


- **Applies to**: Windows 10 / 11 (64-bit) (Windows 11 English UI)
- **Software version**: DeepChat 1.1.4
- **Screenshots updated**: 2026-11
- **Time needed**: about 10 minutes, including downloading the 312 MB installer
- **You will need**: a QCode API key starting with `cr_`


Before you start, make sure you have these ready:


- A 64-bit Windows 10 or 11 PC that can open `deepchatai.cn`, `github.com`, and `api.qcode.cc`
- About 1.5 GB of available disk space (312 MB for the installer plus 1.1 GB for DeepChat, installed into your own user folder)
- A QCode API key copied from your dashboard (`cr_xxxxxxxx`). No account yet? [Sign up](https://qcode.cc/en/auth/sign-up?utm_source=docs.qcode.cc&utm_medium=tut_before&utm_campaign=knowledge_base&src=docs-ide-deepchat-before) first, pick a plan on the [pricing page](https://qcode.cc/en/pricing?utm_source=docs.qcode.cc&utm_medium=tut_before&utm_campaign=knowledge_base&src=docs-ide-deepchat-before), then create the key in your [Dashboard](https://qcode.cc/en/dashboard).
- Note: on a fresh Windows 11 PC without Microsoft Visual C++ Redistributable (x64), the installer prompts to install that runtime once before DeepChat Setup starts


## Download and install DeepChat

DeepChat distributes Windows builds as `.exe` installers via GitHub Releases (there is no Microsoft Store or `winget` package yet). Pick the tab that matches your network:

**Other regions**

Open [deepchatai.cn/#/download](https://deepchatai.cn/#/download) and click **Windows x64** (312 MB) for standard Intel/AMD PCs, or **Windows ARM64** (266 MB) for Windows on ARM. You can also download `DeepChat-1.1.4-windows-x64.exe` directly from [GitHub Releases](https://github.com/ThinkInAIXYZ/deepchat/releases). Do not download `.blockmap` files.


**Mainland China**

Open [deepchatai.cn/#/download](https://deepchatai.cn/#/download) and wait a second or two for the download buttons to appear, or open [GitHub Releases](https://github.com/ThinkInAIXYZ/deepchat/releases) directly. Click **Windows x64** (`DeepChat-1.1.4-windows-x64.exe`, 312 MB) for ordinary Intel/AMD PCs, or **Windows ARM64** for ARM PCs. Because the installer downloads from GitHub Releases, the connection may be slow or stall from mainland China — if it pauses, retry later or switch networks. Ignore `.blockmap` files (those are update checksums, not installers).




### Download the Windows installer

Open [https://deepchatai.cn/#/download](https://deepchatai.cn/#/download) in your browser and wait a moment for the page to render. Click ① **Windows x64** (`312 MB`) for a standard Intel or AMD PC; only click ② **Windows ARM64** (`266 MB`) if you use a Windows-on-ARM device.

![Open the DeepChat download page (https://deepchatai.cn/#/download) and click Windows x64 (312 MB) for standard Intel/AMD PCs, or Windows ARM64 for ARM devices](https://docs.qcode.cc/static/images/tut/deepchat/202611/en/01-download-page.webp)

**Stuck? Edge asks to translate the page, or the download is slow**

The page shows English buttons on English Windows, but its browser tab title still contains Chinese text, so Edge may pop up a translation prompt — simply close that popup. The button pulls `DeepChat-1.1.4-windows-x64.exe` from GitHub Releases; if it stalls, cancel and retry from [GitHub Releases](https://github.com/ThinkInAIXYZ/deepchat/releases) and make sure you pick the `.exe` file, not `.blockmap`.



Done right you will see: `DeepChat-1.1.4-windows-x64.exe` (about 312 MB) finishes downloading into your **Downloads** folder.




### Choose installation options

Double-click `DeepChat-1.1.4-windows-x64.exe` to start **DeepChat Setup**. Keep ① **Only for me** selected — this installs for your user account without needing administrator rights — then click ② **Next >**.

![Run DeepChat-1.1.4-windows-x64.exe, keep the default Only for me option (no administrator or UAC prompt needed), and click Next](https://docs.qcode.cc/static/images/tut/deepchat/202611/en/02-install-options.webp)

**Stuck? Windows SmartScreen or a Visual C++ prompt appears first**

Two prompts can appear before the wizard:

- **Windows Defender SmartScreen**: the installer is unsigned, so Windows may show **Windows protected your PC**. Click **More info**, then **Run anyway**.
- **Visual C++ Redistributable**: on a clean Windows 11 PC without the runtime, a dialog says `NOTE: DeepChat requires 'Microsoft Visual C++ Redistributable' (x64) to function properly. Download and install now?`. Click **Yes**, agree and click **Install** (this one-time step triggers a UAC prompt), then close it when finished so **DeepChat Setup** can continue.



Done right you will see: ① **Only for me** is selected and clicking ② **Next >** moves to the install location screen with no UAC prompt.




### Confirm the install folder and finish setup

On **Choose Install Location**, keep the default per-user path (`AppData\Local\Programs\DeepChat`, about 1.1 GB required) and click ① **Install**. When the final screen appears, leave **Run DeepChat** checked and click **Finish**.

![Confirm the default per-user destination folder (about 1.1 GB required) and click Install](https://docs.qcode.cc/static/images/tut/deepchat/202611/en/03-install-location.webp)


Done right you will see: the progress bar finishes, the **Finish** button appears, and DeepChat launches automatically after you click **Finish**.




### Skip the first-launch Setup Guide

On first launch, DeepChat shows **Welcome to DeepChat Agent** with an 8-step **SETUP GUIDE (1/8)** spotlight overlay. Click ① **Skip All** to jump straight into the main chat window.

![DeepChat launches after installation with an 8-step Setup Guide spotlight; click Skip All to enter the main window](https://docs.qcode.cc/static/images/tut/deepchat/202611/en/04-first-launch-guide.webp)


Done right you will see: the dark spotlight overlay disappears and the main DeepChat workspace opens.




### Expand the sidebar and open Settings

DeepChat starts with its left sidebar collapsed, hiding the **Settings** gear button behind **Plugins**, **Theme · System**, and **Expand sidebar**. First click the bottom-left sidebar button at ① (**Expand sidebar**, which turns into **Collapse sidebar** once open), then click the ② **Settings** gear icon right next to it.

![The sidebar starts collapsed: expand it from the bottom-left icon, then click the Settings gear icon](https://docs.qcode.cc/static/images/tut/deepchat/202611/en/05-open-settings.webp)

**Stuck? I cannot find the Settings gear icon**

When the left sidebar is collapsed, the gear icon is hidden. Click the bottom-left **Expand sidebar** icon first; once the sidebar widens, the **Settings** gear appears at the bottom right of the sidebar.



Done right you will see: a separate **Settings** window opens, with a **← Back to chat** button at the top left.




### Open Provider Center and add a custom provider

In the **Settings** window, click ① **Providers** under **Agent Settings** on the left. In the middle **Provider Center** column, click ② **+ Add Custom Provider**.

![In the Settings window, open Providers under Agent Settings, then click + Add Custom Provider in Provider Center](https://docs.qcode.cc/static/images/tut/deepchat/202611/en/06-provider-center.webp)


Done right you will see: the **Add Custom Provider** form opens inline in the right-hand pane of the Settings window.




### Name the provider and select OpenAI Completions

In the inline **Add Custom Provider** form, type `QCode` into ① **Name**. Then open the **API Type** dropdown and pick ② **OpenAI Completions**.

![Enter QCode as the Name and open API Type to choose OpenAI Completions (choose Anthropic instead if you want Claude models)](https://docs.qcode.cc/static/images/tut/deepchat/202611/en/07-provider-api-type.webp)

**Stuck? Which API Type should I pick?**

Pick **OpenAI Completions** (not **OpenAI**) for GPT and Chinese models such as `gpt-6.2-sol`, `glm-5.3`, `kimi-k3`, and `deepseek-v4-pro`. If you want Claude models like `claude-opus-5-5`, add a second custom provider using **Anthropic** (see **Add Claude models via Anthropic** below).



Done right you will see: ① shows `QCode` and ② **OpenAI Completions** is selected as the API Type.




### Enter the Base URL and API Key, then connect

In ① **API Base URL**, enter `https://api.qcode.cc/openai/v1` — DeepChat automatically previews `API Example: https://api.qcode.cc/openai/v1/chat/completions` underneath, so do not add `/chat/completions`. Paste your `cr_` key into **API Key** (it masks as dots), then click ② **Connect and load models**.

![Enter https://api.qcode.cc/openai/v1 as the API Base URL (DeepChat previews /chat/completions underneath), paste your cr_ API Key, and click Connect and load models](https://docs.qcode.cc/static/images/tut/deepchat/202611/en/08-provider-url-and-key.webp)

**Need an API Key?**

If you do not have a QCode account yet, [Sign up](https://qcode.cc/en/auth/sign-up?utm_source=docs.qcode.cc&utm_medium=tut_key_step&utm_campaign=knowledge_base&src=docs-ide-deepchat-key) first. Purchase a pack on the [Pay-as-you-go packs](https://qcode.cc/en/pay-as-you-go?utm_source=docs.qcode.cc&utm_medium=tut_key_step&utm_campaign=knowledge_base&src=docs-ide-deepchat-key) page if you prefer not to subscribe to a monthly plan, then create your key in the [Dashboard](https://qcode.cc/en/dashboard) and copy the full `cr_` string. For available payment options, including credit and debit cards, see [Payment methods](/docs/reference/payment-methods).


**Stuck? Connection fails with 401 or 404**

A `404` error usually means the URL is missing `/v1` or already includes `/chat/completions` — make sure **API Base URL** is exactly `https://api.qcode.cc/openai/v1`. A `401` error means the key was copied with extra spaces, quotes, or line breaks, or is invalid; copy the full `cr_` key again from your [Dashboard](https://qcode.cc/en/dashboard).



Done right you will see: DeepChat connects and displays **QCode is connected**, showing that it loaded QCode’s model list along with **Start chatting** and **View models** buttons.




### Confirm QCode and gpt-6.2-sol are enabled

Click **View models** to open the **QCode** provider details. Verify the green **Last test passed** badge and that ① **Enable** at the top right is switched on. Type `gpt-6.2-sol` in **Search models...** and confirm ② its toggle is turned on.

![After connecting, QCode shows Last test passed with Enable turned on and loads all 20 QCode models; search gpt-6.2-sol to confirm its toggle is enabled](https://docs.qcode.cc/static/images/tut/deepchat/202611/en/09-provider-connected-models.webp)

**Stuck? You clicked Save without testing, or a model is missing**

If you clicked **Save without testing**, the provider starts turned off. Switch ① **Enable** on at the top right (which also fetches models automatically), or click **+ Add Model**, enter `gpt-6.2-sol` in both **Model Name** and **Model ID**, and click **Save configuration**.



Done right you will see: **Last test passed** in green, ① **Enable** switched on, and ② `gpt-6.2-sol` switched on inside **Model List**.




### Return to chat and pick gpt-6.2-sol

Click **← Back to chat** in the top-left corner of the Settings window. DeepChat may default to another preselected model (such as `gpt-6.2-astra`), so click ① the model selector button below the chat input box, then click ② `gpt-6.2-sol` under the **QCODE** group.

![Click Back to chat, open the model selector below the input box, and choose gpt-6.2-sol under QCODE](https://docs.qcode.cc/static/images/tut/deepchat/202611/en/10-select-model.webp)


Done right you will see: the model selector button at ① updates to show `gpt-6.2-sol`.




### Send a test message

Type a short prompt such as `Introduce yourself in one short sentence.` and press **Enter**. Your message appears at ①, and within a few seconds ② `gpt-6.2-sol` streams a complete reply.

![Send a test message and receive a live reply from gpt-6.2-sol within seconds, confirming QCode is working](https://docs.qcode.cc/static/images/tut/deepchat/202611/en/11-first-reply.webp)


Done right you will see: ① your sent message and ② a live reply from `gpt-6.2-sol` with no red error banner.




## Verify it works

Your setup is complete when all three of these are true:

- In **Settings → Providers → Provider Center**, **QCode** shows **Last test passed** with **Enable** turned on and `gpt-6.2-sol` enabled in **Model List**.
- In the main chat window, `gpt-6.2-sol` is selectable under the **QCODE** group in the model picker.
- Sending a test prompt returns a live reply within seconds, and the request appears in your [Dashboard](https://qcode.cc/en/dashboard) usage records.


### All done

DeepChat is now connected to QCode on Windows 11. From here you can:

- [Endpoints and API paths](/docs/getting-started/endpoints-and-api-paths) — compare the OpenAI and Anthropic paths and regional hosts
- [Model selection](/docs/usage/model-selection) — browse available GPT, Claude, and other model IDs
- [Billing](/docs/reference/billing) — understand how token usage is calculated
- [Troubleshooting](/docs/reference/troubleshooting)
- Compare plans on the [pricing page](https://qcode.cc/en/pricing?utm_source=docs.qcode.cc&utm_medium=tut_done&utm_campaign=knowledge_base&src=docs-ide-deepchat-done)
- Go back to your [Dashboard](https://qcode.cc/en/dashboard) to check usage and balance


## Common errors

| Error message | Why | What to do |
|---|---|---|
| `NOTE: DeepChat requires 'Microsoft Visual C++ Redistributable' (x64) to function properly. Download and install now?` | The 64-bit Visual C++ runtime is not yet installed on this Windows PC | Click **Yes**, complete the Microsoft Visual C++ 2015–2022 Redistributable (x64) installer (approve the UAC prompt), and close it; DeepChat Setup continues right after |
| `Authentication failed, API Key is invalid or expired` | The `cr_` key was pasted with extra spaces or line breaks, comes from another service, or has expired | Copy the full `cr_` key again from your [Dashboard](https://qcode.cc/en/dashboard) and paste it cleanly into **API Key**; if it expired, follow the renewal rule in **FAQ** below |
| `model_not_available_on_endpoint` | A Claude model (such as `claude-opus-5-5`) was called through an `OpenAI Completions` provider (`https://api.qcode.cc/openai/v1`) | Add a separate custom provider with **API Type** set to `Anthropic` and **API Base URL** set to `https://api.qcode.cc/api` (see **Add Claude models via Anthropic**) |
| `claude-sonnet-4-5-20250929` (probe fails when clicking **Connect and load models** on an `Anthropic` provider) | DeepChat probes new `Anthropic` providers with a built-in model name instead of your target QCode Claude model | Click **Save without testing**, add `claude-opus-5-5` manually via **+ Add Model**, and switch **Enable** on |
| `404` | **API Base URL** is missing `/v1` (for `OpenAI Completions`) or has an extra `/chat/completions` or `/v1` appended (for `Anthropic`) | Use exactly `https://api.qcode.cc/openai/v1` for `OpenAI Completions` and `https://api.qcode.cc/api` for `Anthropic`, with no trailing path |

## FAQ

### What happens when my QCode key expires?

It depends on how you purchased it:

- **Personal plan**: an expired key cannot be renewed. Activate a new plan, create a new key in your [Dashboard](https://qcode.cc/en/dashboard), and replace the old key in DeepChat’s QCode provider settings.
- **Pay-as-you-go pack**: purchase another pack and your existing `cr_` key resumes automatically — you do not need to change anything in DeepChat.

### Why should I pick OpenAI Completions instead of OpenAI?

In DeepChat’s **API Type** list, **OpenAI** follows Responses path rules, whereas **OpenAI Completions** sends standard Chat Completions requests to `https://api.qcode.cc/openai/v1` and loads QCode’s model list automatically when you click **Connect and load models**.

### Can I use both GPT and Claude models in DeepChat?

Yes, with the same `cr_` API key. Use two separate custom providers: one **OpenAI Completions** provider (`https://api.qcode.cc/openai/v1`) for GPT and Chinese models, and one **Anthropic** provider (`https://api.qcode.cc/api`) for Claude models (see **Add Claude models via Anthropic** below).

### How do I change DeepChat’s interface language?

Expand the left sidebar, open **Settings**, go to **Setup → Display**, and select your preferred language from the **Language** dropdown.

### What if I clicked Save without testing or do not see a model in the picker?

A provider saved with **Save without testing** starts disabled. Open **Settings → Providers → Provider Center**, click **QCode**, and turn on the **Enable** toggle in the top right (for **OpenAI Completions**, this also fetches models). If you need a specific model ID, click **+ Add Model**, fill **Model Name** and **Model ID**, click **Save configuration**, and confirm its toggle is enabled in **Model List**.

## Add Claude models via Anthropic

QCode serves Claude models only on the Anthropic Messages endpoint, so keep Claude in its own custom provider alongside your `OpenAI Completions` provider:

1. Open **Settings → Providers → Provider Center** and click **+ Add Custom Provider**.
2. Fill **Name** with `QCode Anthropic`, set **API Type** to `Anthropic`, set **API Base URL** to `https://api.qcode.cc/api` (do **not** add `/v1` — DeepChat appends `/v1/messages` automatically), and paste the same `cr_` API key.
3. Click **Save without testing** instead of **Connect and load models** (DeepChat’s built-in Anthropic probe uses `claude-sonnet-4-5-20250929` and stays on the form).
4. In the saved **QCode Anthropic** provider page, click **+ Add Model**, enter `claude-opus-5-5` for both **Model Name** and **Model ID**, and click **Save configuration**.
5. Switch on the **Enable** toggle at the top right, click **← Back to chat**, and pick `claude-opus-5-5` from the model selector.

## Access points and available models

| Protocol (API Type) | API Base URL | Example models on QCode |
|---|---|---|
| `OpenAI Completions` | `https://api.qcode.cc/openai/v1` | `gpt-6.2-sol`, `gpt-6.2-astra`, `glm-5.3`, `kimi-k3`, `deepseek-v4-pro` |
| `Anthropic` | `https://api.qcode.cc/api` | `claude-opus-5-5` |

If you want to route through the Asia or North America access point, replace `api.qcode.cc` with `asia.qcode.cc` or `us.qcode.cc` while keeping the exact same path (`/openai/v1` or `/api`). All three domains offer the same capabilities and differ only in network routing.

## Related docs

- [Endpoints and API paths](/docs/getting-started/endpoints-and-api-paths) — OpenAI and Anthropic paths and regional domains
- [Model selection](/docs/usage/model-selection) — current GPT, Claude, and other model IDs on QCode
- [Payment methods](/docs/reference/payment-methods) — available ways to pay for plans and pay-as-you-go packs
- [Billing](/docs/reference/billing) — how token-based billing works
- [Troubleshooting](/docs/reference/troubleshooting) — general connection and authentication checks