Possible to connect multiple Agents to the same memory?

Questions
1

Hiya,

Suppose I have three separate flows, one with Llama, one with Mistral, and one with DeepSeek. I like to have them ‘talk’ to each other. But it would be great if they know what the other responded.

The way I’m imagining this is they are all connected to the same memory.

It doesn’t have to be fully automated, I presume that I need to copy paste the reply from one agent, into the input of another and trigger the reply. Without the shared memory, the question I’d ask would be completely new to it every time, instead I like it to know what the other agent replied and respond to that.

Is this possible?

Cheers.

PS: I’m using n8n locally installed with npm, Ollama with Llama3, Mistral, and DeepSeek, and have AirTable.

2

It looks like your topic is missing some important information. Could you provide the following if applicable.

  • n8n version:
  • Database (default: SQLite):
  • n8n EXECUTIONS_PROCESS setting (default: own, main):
  • Running n8n via (Docker, npm, n8n cloud, desktop app):
  • Operating system:
3

Yes, you can connect multiple AI agents to the same memory using this approache:

  1. Using AirTable as shared memory:
  • Create an AirTable table with columns for timestamp, agent_name, message
  • Before each agent runs, add an AirTable node to fetch the conversation history
  • After each response, append the new message to AirTable
  • Format history into proper context string for each agent
1 Like
4

There is another way where you can your json file to store the transferable data.

5

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.