Asset Name | Asset Type | Description |
|---|---|---|
PineconeRAGConsumption | App connection | Pinecone connection. |
GeminiRAGConsumption | App connection | Gemini connection. |
VectorMatchEntry | Process object | Provides all the information from the Pinecone vector matching. |
Get Gemini Embeddings | Process | Subprocess that creates embedding vectors from input text using Gemini AI. |
Pinecone Upsert Embedding | Process | Subprocess that saves vectors to the Pinecone index. |
Fill Empty Pinecone Index Using Gemini AI | Process | Subprocess that creates vectors using Gemini AI and saves them in an empty Pinecone index. |
Generate Embeddings Model | Process | Subprocess that parses user input into a vector view and matches it with the embedding model. Returns the matching context score and metadata. |
Query Pinecone | Process | Subprocess that searches coincidences from input vectors to vectors in the base Pinecone index specified as input parameter Index_Host_Pinecone. Returns a result with metadata. |
Query LLM With Context (Gemini) | Process | Subprocess that prepares and sends requests to the Gemini LLM with the context. |
Get Embeddings | Process | Subprocess that generates embedding vectors from the user input using Gemini. |
Query LLM with Context using Embeddings Model | Process | Submits a query that is received from the user, converts it into a vector, and uses it to search for similar vectors in a database. The top K matches are retrieved, filtered by a cutoff score, and used to form a context. This context including the original query is passed to a Large Language Model (LLM) to generate and return a comprehensive response. |