AgentWeb launches callable website tasks so AI agents skip the browser
Overview
AgentWeb has launched an infrastructure layer that lets AI agents complete website tasks by calling pre-built actions instead of reading screenshots and clicking buttons.
Each mapped task is exposed as a tool with an input schema, preconditions, an authority boundary and a completion proof. Access is by request at agentweb.us.
Developer Rohan Paul, sharing the launch on X, says AgentWeb completed 47 of 47 live tasks across 15 government portals, 7 insurers and 25 signed-in sites such as Amazon and LinkedIn, with a 1.9-second median and $3 total cost. These figures come from the launch post and have not been independently verified.
Written by AI from the articles below · updated Oct 29, 12:02 PM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
Rohan Paul@rohanpaul_aiXAgentWeb launches callable web tasks for AI agents without a browserAIRohan Paul shares the launch of AgentWeb, which lets AI agents complete website tasks by calling pre-built actions instead of reading screenshots and clicking buttons. He says it completed 47 of 47 live tasks across 15 government portals, 7 insurers and 25 signed-in sites like Amazon and LinkedIn, with a 1.9-second median and $3 total cost. Each mapped task is exposed as a tool with an input schema, preconditions, an authority boundary and a completion proof. The quoted AgentWeb post directs readers to request access at agentweb.us.

Andrey@Andrey__HQXAgentWeb launches web infrastructure for agents, reporting 47/47 live-site task successAIAgentWeb introduces web infrastructure for generalized agents, reporting 47 of 47 real-world tasks completed on live sites. The post gives a 1.9-second median time and $3 total cost, and invites users to request access at agentweb.us.

Heat Trend
Not enough continuous observations to show a trend yet.