n8n Embeddings Lemonade nodeConfigure Embeddings Lemonade in n8n.
Embeddings turn a piece of text into numbers a machine can compare. The n8n Embeddings Lemonade node does that with models hosted on a Lemonade server, and exposes a single parameter, Model. It is a sub-node, so it feeds a root node through the Embeddings port instead of running on its own.
Verified Trustpilot reviews · AI, automation & growth agency
What does the n8n Embeddings Lemonade node do?
It generates vector embeddings, meaning numerical representations of text, using a model loaded and managed by a Lemonade server. Anything that compares meaning rather than exact words runs on those vectors: semantic search, clustering, similarity matching. The node itself holds no logic and no input of its own. It is a sub-node, a node that attaches to a port of another node instead of sitting in the main flow, and it hands its embeddings to whatever root node asked for them.
The first scenario is the one most people arrive with: a knowledge base you want to question in plain language. A root vector store node receives your documents through its main input, asks this node for the vectors, and stores them. Pick the model once in the Model dropdown and every document goes through the same Lemonade instance, which keeps the stored vectors consistent with the ones you generate later at query time.
The second is retrieval behind an agent. An AI Agent answers a question, a vector store tool looks up the relevant passages first, and those lookups need embeddings generated exactly the same way as the stored ones. Same node, same model, on both sides. Mixing two different embedding models between indexing and querying is the classic reason a search returns nothing useful.
The third is plain similarity work with no chat involved at all: deduplicating incoming support tickets, grouping form answers pulled from Google Sheets, spotting near-identical product descriptions. Embeddings are enough on their own here, and no chat model is involved.
When would you pick something else? Lemonade makes sense when the models run on a server you control and you want the text to stay there. If you would rather call a hosted provider and not run a server, the equivalent embeddings sub-nodes for OpenAI, Google Gemini or Anthropic sit in the same port and swap in without touching the rest of the workflow.
The limits are worth knowing up front. This node has one job and one parameter: it does not chunk your text, it does not store anything, and it never runs on its own. The model list is whatever your Lemonade server currently serves, so an empty dropdown is a server question, not an n8n question. The page describes version 1 of the node, the highest in the catalog; an older workflow may show fewer options.
What do you need to authenticate?
- 01
Create the Lemonade credential
The node reaches a Lemonade server at runtime, so it needs a credential for it. Open the Credentials menu in n8n and create the Lemonade credential, using the connection details of the server that hosts your models. n8n stores it once and every workflow on that instance can reuse it, on n8n Cloud as on a self-hosted instance.
- 02
Select it on the node
Open the Embeddings Lemonade node and pick the credential you just saved in the credential selector at the top of the panel. There is no Authentication dropdown to set here: the node has a single credential type, so choosing the saved entry is the whole step. A node left without a credential cannot reach the server and its model list stays empty.
- 03
Pick a model and wire the port
With the credential attached, the Model dropdown fills with what your Lemonade instance serves. Choose one, then drag the node onto the Embeddings port of the root node that will use it. Nothing runs until that connection exists, since a sub-node is executed by its root node and never by itself.
Where does this node plug in?
n8n splits AI work between a root node, which sits in the main flow and receives items, and sub-nodes that attach to its ports, one type of sub-node per port.
Output (what it plugs into)
- Embeddings
ai_embedding
01Embeddings
This is the only connection the node has, and it is an output: the node offers embeddings and waits for a root node to ask. The port is exposed by vector store root nodes and by the chains that read from them, and it accepts one embeddings sub-node at a time.
Key parameters
- Required: the node has no main input and never executes on its own, so it stays idle until a root node is attached to this port.
Which parameter do you actually set?
The Embeddings Lemonade node has one parameter. For each one: the node as you configure it in n8n, what the parameter changes, and our field notes.
Model
modelWhat you see in n8n
Notes & use cases
One dropdown, and it decides everything the node produces. It names the model that generates the embeddings, chosen from the list of options served by the Lemonade server configured for this node.
Key parameters
- Model: required, and populated from your Lemonade instance rather than typed by hand. Models are loaded and managed on the server, so the list changes when the server does.
Need help automating Embeddings Lemonade with n8n?
A person reads every message.
Common questions
01Is the Embeddings Lemonade node included in n8n?
02What do you need to authenticate before it works?
03What are the limits of the n8n Embeddings Lemonade node?
04What should you plug in to get started?
05n8n or Make for this kind of AI work?
Get our weekly integration tips.
No spam. Unsubscribe anytime.
