It's important that the software services you use are transparent. Although you may not understand the entirety of the technical details, it's crucial to have some knowledge about the product you are contributing to and how it operates.
From day one, we have made our search plugin service's code open source. You can access the first versions here: https://github.com/ReminisApp/websearch-chatgpt-plugin.
Today, I will explain the principles we follow at Keymate to ensure fair access to information and to allow large language models to operate based on facts and common sense, rather than conjecture. Keymate is a mandatory registration (you log in with your email/Google email) search plugin product for ChatGPT. I believe the decisions we've made were based on the principles listed below:
1. We don't scrape search engine results; we are their customers and we pay them. Other search engine plugins may not do this, but we do. We consider any other approach to be software piracy. We respect the technology built by other developers and therefore, we pay to use them. This is why offering a completely free plugin is not sustainable.
2. We don't fully scrape page content links by default and we provide link attribution so that every piece of content is just an overview of the real content. We attribute the source and ensure that large language models redirect you to the given link. We designed and wrote the plugin description in this way, leaving it up to the large language models to decide whether to provide links by default or not.
3. Quality over quantity: We listen to our users and closely monitor the quality of responses and usage. Maintaining quality for 200,000+ users is not easy, but we are working hard to achieve it.
4. The most important principle is our obsession with information sharing. When information is accessed fairly, quickly, and in a usable way, it propels humanity forward. This is why information technologies have significantly boosted human productivity. We thought that even if no one else uses it (the plugin), we need it, so we made it. If we lose time/investment/money, we don't care because it's a contribution to ourselves and humanity.
We created Keymate.AI while working 9-5 jobs in the UK and US, we funded its creation ourselves with a simple goal: to focus on users and their needs as opposed to satisfying investors. The driving force behind our development was the belief that information sharing and processing should be more robust, widespread, and fast.
We try to keep our costs for the users as low as possible, while reinvesting any and all money we make back into improving the products and experience for users. We aren't trying to maximize profits; our goal is to reach as many people as possible and provide a fair and sustainable tool to help our users in the age of AI.
We are a part of startup programs at Amazon, Google, and Microsoft, and we are using the infrastructures of all three companies. We strive to be the bridge between the best search engine provider, the best infrastructure provider, and the best Large Language Model provider.
With all this in mind, we invite you to become a part of our journey to create a sustainable search tool for LLMs that bridges the gaps between platforms.
Unlike other browsing plugins, Keymate uses the official Google Search API and for each request it pays Google accordingly. Keymate.AI pays for all infrastructure and services that are used to create and operate our products.
You can check our terms of service and privacy policy from https://keymate.ai
If you want to learn more about what we think about ad-based free search engines and ad-based web please check out our video below: