Score DeepSeek V4 Pro for Free! Unlimited Use and Claude-Code Integration

Let's start with a conclusion: By following Brother Xing's steps, you'll be able to directly call DeepSeek V4 Pro's free quota from your local tools, no top-ups, no complicated operations, just 5 easy steps!

Free models:

I. Preparation

To snag this freebie, you just need to prepare these two tools/accounts in advance; both are easy to acquire:

1. ZenMux account: Register directly on the official website (zenmux.ai). This is the key platform for connecting to the free version of DeepSeek V4 Pro.
2. cc switch (not mandatory): A local API forwarding/switching tool, its core function is to integrate third-party APIs into your local environment. It also supports model mapping and forwarding, so friends who work with APIs should be familiar with it;

If you haven't installed CC, you can refer to Brother Xing's previous article: Install Claude Code and cc-switch

cc switch is not mandatory; if you have other large model software like Cherry Studio, you can also use it to connect and use DeepSeek V4 Pro.

II. Process

The entire operation logic is very clear: "Register → Create Key → Restrict Models → Configure Tool → Test". No detours, Brother Xing has broken it down clearly for everyone:

Step 1: Download and Install cc switch

First, download cc switch (you can search for it through your usual channels or find reliable resource sites). After installation, don't rush to configure it; wait until you get the API Key later.

Step 2: Register ZenMux Account

Open the ZenMux official website, click the registration button. Brother Xing recommends logging in directly with Google; it takes a few seconds and saves time by avoiding tedious information entry.

Step 3: Create Exclusive API Key

After logging into ZenMux, click on the avatar in the top right corner, find and enter the "PAYG API" page, then click "Create new API Key". Once generated, copy and save it; you'll need it later.

As shown: Click to create API key

Step 4: Restrict Model Scope (Crucial!)

Do NOT skip this step, or you might not be able to call the free models. When creating the API Key, you'll see the "Model Restriction" option. Enter the following free models here:

deepseek/deepseek-v4-flash-free deepseek/deepseek-v4-pro-free GLM4.6V Flash GLM4.7 Flash

Brother Xing reminds: The flash-free version is faster, suitable for lightweight tasks like information retrieval and simple Q&A; the pro-free version is more capable, perfect for complex reasoning, coding, and writing long articles. After filling them in, just click confirm.

Copy the key

III. Configure Cherry Studio

Brother Xing likes to use Cherry Studio to test if the key is working correctly.

1. Add Provider

2. Fill in Key and API Address

Key: The key you just applied for API Address: See screenshot

3. Add Model

Click to get model list, search for 'free'

4. Test Conversation

Its reply: I am the latest version of DeepSeek's model, but not DeepSeek-v4.

IV. Configure cc switch

Open the installed cc switch, create a new custom configuration, and fill in the following information:

API Key: Paste the Key you just created in ZenMux; Model Mapping: Also fill in "deepseek/deepseek-v4-flash-free" or "deepseek/deepseek-v4-pro-free", matching the models restricted earlier.

1. Add Configuration

Enter CC-Switch, click the plus sign

2. Edit Provider

As shown, mandatory fields

API Key: The one you applied for Request Address: https://zenmux.ai/api/anthropic Model Name: Fill in deepseek/deepseek-v4-pro-free for all

3. Verify Configuration Success

Don't rush to use it after configuration. First, initiate a simple request in your local tool. If it returns results normally, it means you have successfully integrated the free version of DeepSeek V4 Pro!

Brother Xing personally tested it, as long as the steps are correct, it usually works on the first try.

V. Summary

Although it's a freebie, pay attention to these points when using it to avoid pitfalls:

1. Regarding stability: After all, it's a free model, so during peak hours, responses might be slow, queued, or even temporarily unavailable, which is normal. Brother Xing suggests that if you encounter this, either switch to the flash-free model or wait 10 minutes and try again;
2. Regarding model capability: While the pro-free version is sufficient, it has slight limitations in context length and reasoning depth compared to the paid version. If you're doing long text processing or high-intensity development, remember to segment your content for better results;
3. Regarding security: NEVER upload your API Key in plain text to public repositories, and don't share it casually with others to prevent misuse and account issues.
4. Free versions will have token quantity limits!

Overall, this free DeepSeek V4 Pro offering is incredibly cost-effective. Whether for personal daily use or for developers testing APIs, it can save a lot of money.