Overview
The Pinecone provider connects LangChat to Pinecone for vector search and document storage.
Features
- ✅ Vector Search - Find similar documents
- ✅ Automatic Embeddings - OpenAI embeddings included
- ✅ Fast Search - Millisecond query times
- ✅ Scalable - Handles millions of documents
Basic Usage
from langchat.vector_db import Pinecone
vector_db = Pinecone(
api_key="pcsk-...",
index_name="your-index"
)
Configuration
vector_db = Pinecone(
api_key="pcsk-...", # Required
index_name="your-index", # Required
embedding_model="text-embedding-3-large" # Optional
)
Make sure your Pinecone index exists before using it.
Embedding Models
text-embedding-3-large (Recommended)
vector_db = Pinecone(
api_key="...",
index_name="...",
embedding_model="text-embedding-3-large"
)
- High quality search
- Complex queries
- Production use
text-embedding-3-small
vector_db = Pinecone(
api_key="...",
index_name="...",
embedding_model="text-embedding-3-small"
)
- Faster queries
- Lower cost
- Simple use cases
Using with LangChat
from langchat import LangChat
from langchat.llm import OpenAI
from langchat.vector_db import Pinecone
from langchat.database import Supabase
llm = OpenAI(api_key="sk-...", model="gpt-4o-mini")
vector_db = Pinecone(api_key="...", index_name="...")
db = Supabase(url="https://...", key="...")
ai = LangChat(llm=llm, vector_db=vector_db, db=db)
Getting a Retriever
Get a retriever for custom search:
# Get retriever
retriever = vector_db.get_retriever(k=10) # Retrieve 10 documents
# Use in custom code
docs = retriever.invoke("your query")
Best Practices
1. Create Index First
Create your Pinecone index before using:
# In Pinecone console or API
# Create index with correct dimensions:
# - text-embedding-3-large: 3072 dimensions
# - text-embedding-3-small: 1536 dimensions
2. Use Appropriate Embedding Model
# ✅ Good: Match model to needs
embedding_model="text-embedding-3-large" # High quality
# ✅ Also good: Faster option
embedding_model="text-embedding-3-small" # Faster
Troubleshooting
Index Not Found
Solution: Create index in Pinecone console first.
Dimension Mismatch
Solution: Ensure index dimensions match embedding model:
text-embedding-3-large: 3072 dimensionstext-embedding-3-small: 1536 dimensions
Next Steps
Built with ❤️ by NeuroBrain