AI is reshaping the search industry chain. Take the two major giants, for example: Google discontinued its Web Search API back in 2016, and Bing, coincidentally, also shut down its Search API this August. Instead, they now offer AI-enhanced search interfaces. Microsoft hopes users will switch to accessing Bing's search functionality through Azure's AI Agent framework. After all, search is bound to be an essential component of future AI Agent workflows. For instance, OpenAI provides a Web Search interface priced at $10 per 1,000 queries (plus the corresponding token costs), which can be integrated into the AgentKit framework. I’ve noticed that there are still products offering search interfaces, such as Brave, Kagi, Exa, and Perplexity, each with its own strengths and unique features. Several of these are relatively new startups, and they’ve deeply integrated AI into their platforms, enabling more advanced search capabilities like semantic search. These serve as an upgraded version of the old-school "keyword" search, allowing for higher pricing. However, I’m a bit concerned about their ability to index the entire web, which might not be as robust as Google’s. The pricing generally ranges from $5 to $20 per 1,000 queries. Hopefully, we’ll see more competitive new products emerge.