Hi everyone I am Shuhao, co-founder of TinyFish, the team building AgentQL. AgentQL is an AI-powered semantic framework designed to enable AI agents to seamlessly interact with the web, using natural language queries.
We built AgentQL to solve the problem of fragile web automation scripts breaking due to UI changes. Our goal was to create web infrastructure tailored for agentic interaction that can handle small but breaking changes.
AgentQL can find elements and data anywhere on a site using natural language selectors. Unlike fragile XPath or DOM selectors, these queries can find the same information regardless of how the page’s structure might change over time.
Because AgentQL uses AI to infer which elements to return, the same query will work no matter how much the target site may change over time.
We're looking for feedback from the HN community as we develop AgentQL further. You can check it out at Agentql.com. Happy to answer any questions! What do you think about the potential for AI agents to interact with the web in this way?
10 comments
[ 2.8 ms ] story [ 34.8 ms ] threadWe built AgentQL to solve the problem of fragile web automation scripts breaking due to UI changes. Our goal was to create web infrastructure tailored for agentic interaction that can handle small but breaking changes.
AgentQL can find elements and data anywhere on a site using natural language selectors. Unlike fragile XPath or DOM selectors, these queries can find the same information regardless of how the page’s structure might change over time.
Because AgentQL uses AI to infer which elements to return, the same query will work no matter how much the target site may change over time.
We're looking for feedback from the HN community as we develop AgentQL further. You can check it out at Agentql.com. Happy to answer any questions! What do you think about the potential for AI agents to interact with the web in this way?