Describe the problem/error/question
I searched for langchain and I couldn’t find it, pls help
What is the error message (if any)?
Please share your workflow
(Select the nodes on your canvas and use the keyboard shortcuts CMD+C/CTRL+C and CMD+V/CTRL+V to copy and paste the workflow.)
- n8n version: 2.27.5
- Database (default: SQLite):
- n8n EXECUTIONS_PROCESS setting (default: own, main):
- Running n8n via (Docker, npm, n8n cloud, desktop app): Docker
- Operating system: Linux
@Solomon_Ezeigwe
The LangChain nodes have not been removed, but they have been integrated natively into n8n’s core as part of the broader AI suite. You will no longer find a single “LangChain node”; instead, the functionality is split across various AI nodes.
Instead of searching for “LangChain,” look for the AI category in the node selector. You can find the same (and expanded) functionality by searching for:
- AI Agent: The primary node for creating AI-driven workflows.
- Chain: For creating specific sequences of AI actions.
- AI Memory / Embeddings / Vector Store: These are now sub-nodes that you connect to an AI Agent or Chain node.
Since you updated and are using Docker, a common cause for these nodes “disappearing” or causing errors is a conflict with an old community package.
- Check this: Go to Settings →→ Community Nodes.
- Action: If you see
@n8n/n8n-nodes-langchain listed there, uninstall it.
- Why? Because LangChain is now built-in, having the old community version installed creates a conflict that can hide the native AI nodes from the editor.
- Restart: After uninstalling, restart your Docker container.
If the nodes still don’t appear after checking the community nodes:
-Clear Browser Cache: Sometimes the UI doesn’t update the node list immediately after an update.
-Force Fresh Pull: Ensure you are using the latest image by running:
docker pull n8nio/n8n:latest
Then recreate your container.
Nice, but the node I want hasn’t been added yet, I’m looking for a node similar to text classifier but has a memory, it will be able to classify text but with an in built memory that’s why I need the langchain node
In the current version of n8n, you no longer use a single “LangChain node” to achieve this. Instead, you compose the functionality by connecting different AI sub-nodes to a central “brain.”
To build a Text Classifier with Memory, you should use the AI Agent node. This is the direct successor to the complex LangChain chains.
Instead of one node, you will set up a small cluster of nodes like this:
-
The “Brain” (Central Node):
- Add the AI Agent node.
- The Classification Part: In the System Prompt (or Instructions) of the AI Agent, define the classification rules.
- Example: “You are a text classifier. Your job is to categorize user input into ‘Technical Support’, ‘Billing’, or ‘General Inquiry’. Always respond with only the category name.”
-
The “Intelligence” (Attached to Agent):
- Connect a Chat Model node (e.g., OpenAI Chat Model, Anthropic Chat Model) to the AI Agent. This provides the actual LLM power to do the classifying.
-
The “Memory” (Attached to Agent):
- Connect a Memory node (e.g., Window Buffer Memory or Redis Chat Memory) to the AI Agent.
- Why this works: Because the Memory node is attached, the AI Agent will remember previous classifications and the context of the conversation, allowing it to classify the current text based on what happened in previous messages.
Thanks,
I’ve thought of that and it has a huge bump, it’ll need a code node or a Structured Output Parser sub-node and a switch node to route each decision to each flow; but there’s no problem; but I can mange that, i just though of having all in one node
Thanks, once again