Parallel says its new Parallel Search Turbo is "the fastest and most affordable way to bring the web to your AI," pitching it as a search API for AI agents with roughly 200ms median latency and pricing of $1 per 1,000 searches. In a quote-post on X, Parag Agrawal framed that pitch as a change in what the web needs to do for software, writing that "before agents, there was no reason to optimize web search to the extreme."