Automation library · AI assistants

Reduce LLM Costs with Semantic Caching and Redis Vector Store

Started by someone opening a chat with it. It calls services over plain HTTP rather than through connectors. It puts OpenAI behind a retrieval step, so answers are drawn from documents you supply rather than from the model alone. 19 steps, 1 branch: a working afternoon to understand properly.

MediumCalls a model: OpenAIRetrieval

Download free
19 stepsMedium: a working afternoon
1Branch: places the flow splits
0Loops or waits
0Services by connector: it calls them over plain HTTP

The flow

Starts it 2Model 15Logic 2

The steps in the order they run, from what starts it.

  1. 1OpenAI Chat ModelLm Chat Open AiModel
  2. 2Redis Chat MemoryMemory Redis ChatModel
  3. 3When chat message receivedChat triggerStarts it
  4. 4Recursive Character Text SplitterText Splitter Recursive Character Text SplitterModel
  5. 5Embeddings HuggingFace InferenceEmbeddings Hugging Face InferenceModel
  6. 6Embeddings HuggingFace Inference1Embeddings Hugging Face InferenceModel
  7. 7When clicking ‘Execute workflow’Manual triggerStarts it
  8. 8Use Huggingface for embeddingsEmbeddings Hugging Face InferenceModel
  9. 9Recursive Character Text Splitter1Text Splitter Recursive Character Text SplitterModel
  10. 10LLM AgentAgentModel
  11. 11Check for similar promptsVector Store RedisModel
  12. 12Add response as metadataDocument Default Data LoaderModel
  13. 13Store entry in cacheVector Store RedisModel
  14. 14Initialize Redis storeVector Store RedisModel
  15. 15Process sample dataDocument Default Data LoaderModel
  16. 16Analyze results from storeCodeLogic
  17. 17Respond to Chat (from LLM)ChatModel
  18. 18Is this a cache hit?IfLogic
  19. 19Respond to Chat (from semantic cache)ChatModel

How it's built

Starts on
Someone opening a chat with it
Steps
19, with 1 branch
Size
Medium: a working afternoon
Connects to
Nothing by connector; it calls services over plain HTTP
Model
OpenAI
Retrieval
Answers are drawn from documents you supply, not from the model alone

Questions

What does it need to run?

An n8n account, and a key for OpenAI. You connect your own accounts after importing it; none are included.

What starts it?

It starts on someone opening a chat with it.

Does it use AI?

Yes. It calls OpenAI. Its answers draw on documents you give it, not on the model alone. Model calls are billed by the provider, not by n8n.

How long does it take to set up?

Plan an afternoon. 19 steps is enough that you should read each one before running it.

How do I import it?

Copy the workflow, open a new workflow in n8n and paste. The steps appear on the canvas, ready for your own accounts.

Before you run it

Nothing here has been tested by us. It is somebody's working automation, read and described from its own steps. Treat the connections it wants as a list of accounts it will touch, and read the 19 steps before you point it at anything that matters.