Langchain
LangChain is a framework for developing applications powered by large language models (LLMs). SurrealDB is an excellent database choice for LangChain projects primarily because of its multi-model capabilities, which streamline data management by requiring only a single database.
This unified system adeptly handles structured and unstructured data, incorporating vector search, graph traversal, relational queries, full-text search, document storage, and time-series data all within one ACID-compliant engine.
For LangChain applications, which often juggle diverse data types for tasks like context retrieval and complex data interactions, SurrealDB’s ability to consolidate these needs into one platform simplifies architecture, reduces latency, and ensures data consistency, making it a highly efficient and powerful backend solution.
In this guide, we’ll walk through how to use SurrealDB as a vector store for LangChain.
Setup
You can run SurrealDB locally or start with a free SurrealDB Cloud account.
For local, two options:
Install SurrealDB and run SurrealDB. Run in-memory with:
surreal start -u root -p root
Run with Docker.
docker run --rm --pull always -p 8000:8000 surrealdb/surrealdb:latest start
Install dependencies
pip install -U langchain-surrealdb langchain_ollama surrealdb
poetry add langchain-surrealdb langchain_ollama surrealdb
uv add --upgrade langchain-surrealdb langchain_ollama surrealdb
surrealdb → SurrealDB Python SDKlangchain-surrealdb → houses SurrealDBVectorStorelangchain_ollama, langchain-openai (or HF, Cohere, etc.) → embeddings
Quick start
Create a vector store, and documents with embeddings, and do a similarity search.
from langchain_core.documents import Document
from langchain_surrealdb.vectorstores import SurrealDBVectorStore
from langchain_ollama import OllamaEmbeddings
from surrealdb import Surreal
conn = Surreal("ws://localhost:8000/rpc")
conn.signin({"username": "root", "password": "secret"})
conn.use("langchain", "demo")
vector_store = SurrealDBVectorStore(OllamaEmbeddings(model="llama3.2"), conn)
doc_1 = Document(page_content="foo", metadata={"source": "https://surrealdb.com"})
doc_2 = Document(page_content="SurrealDB", metadata={"source": "https://surrealdb.com"})
vector_store.add_documents(documents=[doc_1, doc_2], ids=["1", "2"])
results = vector_store.similarity_search_with_score(
query="surreal", k=1, custom_filter={"source": "https://surrealdb.com"}
)
for doc, score in results:
print(f"* [SIM={score:3f}] {doc.page_content} [{doc.metadata}]")
Under the hood the helper will:
- Create table
documents (if it doesn’t exist). - Add an M-Tree index with the correct dimensionality.
- Insert each text with its freshly generated embedding.
Similarity search
query = "How do I enable vector search in SurrealDB?"
docs = vector_store.similarity_search(
query=query, k=1, custom_filter={"source": "https://surrealdb.com"}
)
for doc in results:
print(f"{doc.page_content} [{doc.metadata}]")
The Vector Search feature of SurrealDB... [{'source': 'https://surrealdb.com'}]
If you want to get the score with the results, use similarity_search_with_score instead.
You can also transform the vector store into a retriever for easier usage in your chains.
query = "How do I enable vector search in SurrealDB?"
docs = vector_store.similarity_search(
retriever = vector_store.as_retriever(
search_type="mmr", search_kwargs={"k": 1, "lambda_mult": 0.5}
)
retriever.invoke(query)
[Document(id='4', metadata={'source': 'https://surrealdb.com'}, page_content='The Vector Search feature of SurrealDB...')]
Next steps
Now that you have a basic understanding of how to use SurrealDB with LangChain, let’s explore some additional resources to help you dive deeper and build more sophisticated applications.
To help you get started quickly, we provide several example implementations:
Further reading and resources
For a deeper understanding of the technology stack:
