As everyone knows, DeepSeek has recently skyrocketed in popularity. After its launch, it quickly topped the App Store and Google Play charts. However, this has been accompanied by numerous issues. Users have found that after asking just two or three questions, they frequently receive the prompt "The server is busy, please try again later," which significantly impacts usability and frustrates users everywhere.
Currently, the most effective workaround is local deployment. However, for some beginners, local deployment can be a cumbersome process. You might follow many tutorials and try numerous times, yet still fail to successfully deploy a large model.
I want to share a less-known, unconventional method I use at work that is incredibly simple. Even beginners or those with no programming experience can quickly learn it.
But please note that this method currently only applies to macOS systems; Windows users are out of luck.
By chance, I discovered that ServBay, which I usually use for development, had been updated. Its new version supports Ollama. Ollama is a tool focused on running large language models (LLMs) locally. It supports well-known AI models such as DeepSeek-Coder, Llama, Solar, Qwen, and more.
So, do you get what I mean? This means that by simply installing ServBay, you can enable these pre-packaged, commonly used AI models with one click, and the response speed is quite good.
Originally, Ollama required a complex process to install and start the service, but through ServBay, it only takes one click to start and install the AI model you need, without worrying about environment variable configuration. Even ordinary users with no development knowledge can use it with a single click. One-click start and stop, multi-threaded fast model download; as long as your macOS can handle it, running multiple large AI models simultaneously is not a problem.
On my computer, the download speed even exceeded 60MB per second, surpassing all other similar tools. See the screenshot for proof.
This way, through ServBay and Ollama, I can deploy DeepSeek locally. Look, it's running smoothly!
See! I've achieved DeepSeek freedom with ServBay~
By the way, the US Congress has recently proposed a new bill that would make downloading DeepSeek a crime, punishable by up to 20 years in prison! However, by deploying locally and using it offline, wouldn't that...? Oh, please forgive my wild imagination, LOL...
Top comments (0)