Building the Internet of Agents: Introducing AGNTCY.org

The Internet wasn't built by a single company. It was built through a collective of thinkers and builders with a shared vision, driven to innovate by open standards. The Internet of Agents will be no different. 

Today, we're announcing the AGNTCY - an open source collective building the critical infrastructure for AI agents to work together. We invite the entire AI and infrastructure community to join us in this town square for collaboration. Your expertise and contributions are essential for creating an open, interoperable foundation that serves everyone. Cisco, LangChain, and Galileo are the initial core maintainers, with Glean and LlamaIndex joining as contributors.

The problem: We need agent infrastructure now

Agentic AI will accelerate all human work. This isn't hyperbole, it's the inevitable next step in computing. We're witnessing a Cambrian explosion of AI agents. While single agents can handle specific tasks, the real power comes when specialized agents collaborate to solve complex problems—just like humans do in the real world. But there's a fundamental gap: we have no standardized infrastructure for these agents to discover, communicate with, and work alongside each other.

Our vision

An open, interoperable Internet of Agents is the best path forward to accelerate innovation and create the most value for all participants–from builders to operators, developers to consumers. Just as the original internet connected computers and the web connected information, the Internet of Agents will connect AI systems across vendor and organizational boundaries, and technical frameworks. Without this foundation, we're essentially trying to build the web without RPCs, HTTP, DNS, or TCP/IP.

Who’s contributing

This collective came to life as a group of visionary core maintainers who each bring critical expertise and a shared commitment to openness. 

LangChain brings deep expertise in both agent orchestration and evaluation. As Harrison Chase, CEO of LangChain, explains, “Multi-agent systems are the future of AI, but we need open standards for both collaboration and rigorous assessment. The AGNTCY creates the foundation for a truly observable and trustworthy AI ecosystem. Our evaluation frameworks ensure agents perform reliably across diverse scenarios, empowering organizations to deploy at scale.” 

Galileo specializes in agent trust and observability. "The agentic future should be open and distributed," says Yash Sheth, Co-Founder of Galileo. “At Galileo, we help organizations build trustworthy agents and scale agentic systems with confidence. We're happy to contribute frameworks and schemas for agent evaluation to this important effort.”

Cisco contributes crucial secure networking infrastructure expertise. "We're at an inflection point in AI where the next wave of innovation will come from agents working together," notes Guillaume De Saint Marc, Vice President of Engineering at Outshift by Cisco. "Building this infrastructure as open source ensures it will serve the entire ecosystem, not just individual vendors."

But this is just the start, join other early members in setting the standards for how agents will communicate.

Critical phases to solve for

In a recent post, we identified four critical phases that need to come together to deliver the best possible multi-agent workflow:

1. Discover: Find and evaluate the best agents for the job.
2. Compose: Connect agents into effective workflows across any framework or vendor.
3. Deploy: Run multi-agent systems at scale, securely.
4. Evaluate: Monitor performance and improve efficiency and efficacy over time.

Without standardized infrastructure sourced from high fidelity open-sourced options, builds are happening in silos and limiting what's possible with multi-agent systems. Our collective approach breaks down these barriers by establishing open specs and reference implementations with architecture built with open-source code that tackles the end-to-end requirements for sourcing, creating, scaling, and optimizing agentic workflows from a single app to thousands.

Where to start with AGNTCY

We’re systematically tackling this entire infrastructure to bring the Internet of Agents to life and you can start reviewing the following specs (with code dropping soon):

Discover

• Agent directory  
  • Registry for publishing and discovering agents
  • Implements reputation tracking and quality metrics
• Open agent schema framework  
  • Defines standard metadata format for describing agent capabilities
  • Includes verification mechanisms for agent providers
  • Specification and reference implementation at github.com/agntcy/oasf

Compose

• Agent connect protocol and SDK
  • Enables standardized collaboration between agents across frameworks
  • Handles message passing, state management, and context sharing
  • Specification available now at github.com/agntcy/acp-spec

What could these look like in action? Our goal is seamless, end-to-end agent-agent collaboration: A developer can use the agent directory to find agents that match their requirements (as defined by OASF), then use the agent connect protocol to enable collaboration regardless of which frameworks the agents were built with.

Expanding the collective

Help us unlock agentic capabilities through open source by joining AGNTCY. There are countless opportunities for members to contribute expertise. Here are just a few ideas and suggestions we’re currently evaluating:

• Agent orchestration and workflow management
• Agent performance monitoring and evaluation
• Domain-specific agent frameworks and OASF extensions
• AI infrastructure and hardware acceleration
• Sample new applications and uses cases

Collaborate with us

The transition to agentic-first computing at scale isn’t a matter of if; it’s a matter of when. Getting the right infrastructure in place now will determine just how transformative these systems will be in the years ahead.

We're building this town hall for collaboration because we believe the best ideas come from diverse perspectives working together. Join us to shape standards, contribute code, share use cases, or simply bring your expertise to the conversation.

Visit agntcy.org to explore our specifications, with code dropping at the end of March. And if you’re interested in becoming a member, email us at hello@agntcy.org (we do have some humans behind the inbox).