Communication between two local agents - how? #914
lrosenthol
started this conversation in
General
Replies: 1 comment 1 reply
-
|
One of the advantages of using HTTPS as the underlying protocol for all the A2A transports and a URL for the Agent identifier is that it provides location independence. Two agents can communicate if they are on the same machine or if they are continents away. The one exception here is the use of .well-known as the agent card discovery mechanism. It is currently not possible to have two agents on the same logical host name. However, use of .well-known is optional. |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
In the diagram on the A2A main page (https://a2a-protocol.org/latest/#a2a-and-mcp-complementary-protocols) it shows a section called "local agents", which I take to mean a series of agents all deployed on the same system. And it shows how two separate systems of agents talk between themselves using A2A.
My question is what if two agents running on the same system - perhaps even a single (disconnected/offline!) device - want to talk to each other. Is A2A the correct solution (using localhost as the endpoint), or is there some other approach that is commonly used?
Beta Was this translation helpful? Give feedback.
All reactions