🔧 A local alternative to Microsoft Copilot in Word

Large language models (LLMs) have been rapidly advancing over the past few years. While bigger and more powerful versions are being developed in the cloud, there’s a growing trend towards running LLMs on your own computer. This shift aims to bring LLMs closer to your device, providing faster and completely private processing of your data. By doing so, you can unlock new possibilities for your writing and reduce costs associated with cloud-based solutions such as Microsoft Copilot in Word.

After conducting research on affordable alternatives at Microsoft AppSource, we discovered several promising Word Add-ins that caught our attention. Add-ins like Ghostwriter, GPT for Word, and Sakti GPT, as well as GenText AI Assistant, offered useful features. However, upon closer inspection, we realized that all of these options are cloud-based, which raised concerns about data privacy and flexibility. We needed a solution that could meet our specific requirements without relying on the cloud. Ultimately, we decided to create a local Word Add-in tailored to our unique needs.

A local Word Add-in is made possible by leveraging the developer mode provided by Microsoft during software development. In this way, before production run, developers can work on the source code and test it locally . This developer mode enables access to local resources that wouldn’t be available if the Add-in were cloud-based at Microsoft AppSource. Today, many powerful LLMs are widely used locally. However, one significant limitation is that they cannot be seamlessly utilized in Microsoft Word.

The local Word Add-in here is called GPTLocalhost, which has been successfully integrated with several popular local LLM Servers or Gateways, such as AnythingLLM , LiteLLM, LM Studio, Ollama, llama.cpp, LocalAI, KoboldCpp, and Xinference. Using the intuitive interface as shown below, you can easily customize your experience by selecting a text range as the input prompt (e.g., the whole document), choosing where to output the generated response (e.g., the end of the document), and providing instructions for the model to follow (e.g., complete the writing). With GPTLocalhost, you can take advantage of these powerful LLM servers to enhance your productivity and creativity within Microsoft Word completely locally.

For maximum flexibility and simplicity, we don’t include LLM servers during installation. You can stick with your existing LLM server or install a new one. As long as it is compatible with the API spec for chat completion defined by OpenAI, it will work with GPTLocalhost. If you encountered any issues connecting to a different LLM server, please email [email protected]. At GPTLocalhost, we are changing the way AI is experienced by leveraging advanced GPT models on your local device. Your data never leaves your device. It stays just with you. We envision a future where AI is accessible, secure, and efficient for everyone — without compromising privacy or performance. For answers to common questions, check our FAQ section or reach out to us with any questions you may have. Start your free trial today! Download GPTLocalhost now and experience its features firsthand.”