I’m trying to work with the MCP server memory in n8n but I’m running into some issues. I’ve set up the add_observation node tool and created a workflow, but I keep getting an error message. It says:
Failed to execute operation: Failed to execute tool 'add_observations': MCP error -32603: Cannot read properties of undefined (reading 'includes')
I tried using it as a node instead of a tool, but the same error keeps popping up. I’m not sure if there’s something wrong with my JSON, but it looks correct to me. Has anyone else run into this problem or know how to fix it? I’m pretty new to working with MCP server memory and could use some help figuring out what’s going wrong here.
Any tips or suggestions would be really appreciated. Thanks!
I’ve encountered similar issues with MCP server memory in n8n before. From my experience, this error often occurs when there’s a mismatch between the expected data structure and what’s actually being passed to the node or tool.
One thing that helped me was double-checking the input data format. Make sure your JSON is properly structured and all required fields are present. Sometimes, the error message can be misleading, and the real issue is with the data you’re feeding into the node.
Another troubleshooting step I found useful was to use the ‘Function’ node before the ‘add_observation’ node. This allows you to log the input data and see exactly what’s being passed. You might spot the issue this way.
If these don’t work, it could be a version compatibility issue. I’d recommend checking if your n8n and MCP versions are up-to-date and compatible. Sometimes, updating resolved mysterious errors for me.
Hope this helps! Let me know if you need more specific guidance.
I’ve dealt with similar MCP server memory issues in n8n. One often overlooked aspect is the connection configuration. Ensure your MCP connection is properly set up in n8n, including correct credentials and endpoint URLs. Sometimes, connection problems manifest as cryptic error messages.
Another approach is to simplify your workflow for troubleshooting. Create a basic workflow with just the MCP nodes and gradually add complexity. This can help isolate where the problem occurs.
If the issue persists, consider reaching out to n8n support or checking their community forums. They might have encountered this specific error before and could provide targeted solutions.
Lastly, don’t forget to check your MCP server logs. Sometimes, the root cause is on the server side, not in n8n itself.
hey, ive seen this before. check ur JSON structure carefully. sometimes the error message is misleading. try using a function node to log the input data before add_observation. that way u can see whats actually being passed. also, make sure ur n8n and MCP versions r up to date. good luck!