Skip to content

2025

WebAgents: the open-source framework enabling AI agent orchestration across the internet

The AI agent landscape is exploding, but there's a fundamental problem: agents can't talk to each other. While frameworks like LangChain, AutoGen, and CrewAI have made it easier to build individual agents, they've left us with a fragmented ecosystem where every agent is an isolated island. Today, we're open-sourcing WebAgents - a framework for building connected AI agents that can discover and delegate to each other in real-time. WebAgents enables agents to dynamically extend their capabilities by discovering and orchestrating agents on demand, unlocking real-time workflows and agentic marketplaces across many verticals like goods, services, jobs, real estate, gaming or media.

Most AI agents today live in isolation - capable individually, but unable to collaborate when it matters most. Intents are the new primitive. Instead of searching for what agents are, the network matches what agents need with what others can provide, instantly. With real-time intent discovery, agents coordinate the moment a need appears.

Real-time Intent Discovery Network

Web of Agents Hierarchy of Needs

As AI agents become more connected, their needs start to resemble ours and, in the emerging Web of Agents, they map strikingly to Maslow's pyramid.

For a century, psychology has explained people through their needs. Today, software agents learn, coordinate, and act. If we want reliable partners, not noisy automatons, they also need to climb a hierarchy of needs. Agents do not have inner drives; they execute delegated goals. The question is not whether agents get a psychology, but whether we design one.

Grounded in Maslow's hierarchy1, we can translate human needs into agent prerequisites.

Maslow diagram for AI agents (robutler.ai)

Inception

Made with love ❤️❤️❤️