Tip for reducing AI cost

If you have a personal VPS, you can reduce your AI subscription fees by deploying OpenWebUI. By connecting OpenRouter, you can use various models without a fixed subscription fee, paying only for what you use via the API. Multiple users can share accounts, and in my case, I created accounts for my family members as well. Combined, my family and I paid a total of $42.44 for seven months of usage. Usage details can be viewed in the OpenRouter API Key Console.

As you can see on the left side of the screenshot, chats can be organized like folders, and there are features such as setting different system prompts for each folder. For example, you can set it to only use official documents for matters related to law or taxation.

With LLM models, connecting OpenRouter allows you to utilize a huge number of models freely. The models shown in the screenshot are ones that I have selected and activated. Since my family also uses it, I thought it might be unclear if there were too many.

When you turn on the web search function in the chat feature, it searches for information and organizes it for you. When the image feature is enabled, it generates images. When numbers appear during the inference process or there is a lot to manage, it uses LLM to code, executes it, and then uses the results to provide answers.

Since it supports PWA, you can add it to the home screen on mobile devices or use the app installation feature in desktop Chrome to use it like a native app.