A standard, open framework for building AI agents is coming from Cisco, LangChain and Galileo

A goal for a purpose Agent will come They are free and comfortable to each other for AI agents from various organizations. However, to reach this point requires interaction and these agents can be built with different LLM, data frames and code.

To achieve the meeting, they must agree on how these agents can do the developers Communicate with each other. This is a difficult job.

Including a group of companies Cisco, Langchain, Llamaindex, Galato and Talk nonsenseNow agntcy, agntcy, an open-source-source collective, to create an unrelated language of an industry standard agent. Agntcy aims to facilitate communication and exchange with another of any AI agent.

Connecting AI agents

“We want to build the internet of agents that accelerate all the work of all the work of all global work, all social interactions, and all social interactions in global scale, global scale, Cisco, Cisco Incubation Lols, Cisco Incubation Lol.

The Pandey likened agntcy’s arrival Transmission Supervisory Protocol / Internet Protocol (TCP / IP) and domain name system (DNS)Helping to help organize the Internet and allow the interaction between different computer systems.

“The road we think about this problem is that the original Internet allows everyone to come together for people and servers and web farms,” ​​he said. “This is the internet of agents, and the only way to do it is to coordinate it.”

Cisco, Langchain and Galileo, Agntcy will act as a contributor with the main protectors, reap and Llamaindex. However, this structure may change because the collective adds more members.

Standardization of an industry that moves fast

AI agents cannot be Islands. To reach their full potential, they must be able to connect with other agents outside an enterprise network. This is where it interacts.

It is quite difficult to determine the standards in the traditional industries; It becomes more difficult for technology such as AI, here and model changes occur in a few months in a few months. However, this is not the first time a standard for a generative AI is offered.

Langchain, one of the main members of AgNTCY, has its own protocol to work with agents based on Langchain. This Agent Protocol, Langchain agents, which began in November last year, allow you to talk to agents generated with Autogenous, Crewai or another framework.

This time announced an anthropic Model context protocol (MCP) in November. This protocol aims to standardize how models and AI tools are connected to information sources. However, if many developers hugged the MCP, this is not yet complete standard.

Age Sheth, Cofounderi of the AI ​​evaluation platform, Galileo said standardization said “critical”.

“Standardization is necessary, in fact, the growing speed of the agent will manage the growing speed. Today, the teams are forced to understand how their infrastructure components will develop from scratch,” Sheth said. “Standardization of many agent systems may be when it has a strong anchor to express and express the performance, accuracy and reliability of these agents, which is intensified only by not identified models.”

Sheth acknowledges that the interaction of AI agents may be complicated. AgnTcy “Developers want to encourage the needs of the need to remove the wheel, which is important to achieve standardization, instead of adapting the API and tools.”

Langchain CEO Harrison Chase, in a separate conversation that creates a standard, especially agents can be easier to build themselves.

“The building agents are already possible. Repeat all of the future, the agents are not something of the future.

To a platform and a language

Pandey imagines aggntcy as more than a number of codes for agents. It will also allow customers to discover agents from various developments working AgnTcy.

“Customers can discover, compile, compile, compile, place, compile, compile, compile, compile, compile, compile and evaluate workplaces,” said Pandey.

Agntcy, in the standard you need to hire more AI players to add new agents to the platform and earn speed. After all, it takes a mass adoption to prevent a large number of competitiveness standards for something to have an industrial standard.

Projects such as agntcy are encountered by a battle. Pandey wants to get as many ideas choices as possible when the team talks to many other industrial players and developing the platform. This will take place.

Meanwhile, enterprises continue to place and even place AI agents. Maybe they will all be able to talk to each other in the future.