🤝

Build agents and AI applications with LangChain, LangGraph, and Oracle AI.

Packages · Quick start · Samples · Documentation · LangChain docs

---

This is the official repository for the LangChain integrations with Oracle Cloud Infrastructure (OCI) and Oracle Database. It provides native LangChain and LangGraph components — in Python and JavaScript/TypeScript — for OCI Generative AI, OCI Data Science, and Oracle AI Vector Search.

📦 Packages

Package	Version	Install	What it provides
langchain-oci		pip install -U langchain-oci	OCI Generative AI & Data Science. Chat models (ChatOCIGenAI, ChatOCIOpenAI, ChatOCIModelDeployment), embeddings (OCIGenAIEmbeddings), agents (create_oci_agent, create_deepagents_agent), guardrails (OCIGuardrails), and multimodal (vision, PDF, video, audio) support across Cohere, Google, Meta, OpenAI, and xAI models on OCI.
langchain-oracledb		pip install -U langchain-oracledb	Oracle AI Vector Search. Vector store (OracleVS), hybrid & text-search retrievers, document loaders (OracleDocLoader), text splitter, in-database embeddings and summaries, semantic cache, and chat message history — all powered by Oracle Database.
langgraph-oracledb		pip install -U langgraph-oracledb	LangGraph persistence on Oracle Database. Checkpointers (OracleSaver, AsyncOracleSaver) for durable graph state and a key/value store (OracleStore, AsyncOracleStore) with optional vector search for long-term agent memory.
@oracle/langchain-oracledb		npm install @oracle/langchain-oracledb	LangChain.js for Oracle Database. TypeScript/JavaScript counterparts of the Vector Search components: OracleVS, OracleDocLoader, OracleEmbeddings, OracleSummary, and OracleTextSplitter.

Note

This project merges and replaces the earlier OCI and Oracle AI Vector Search integrations from langchain-community. All packages assume you have credentials configured for the OCI and/or Oracle Database services you use.

🚀 Quick start

Chat with OCI Generative AI

from langchain_oci import ChatOCIGenAI

llm = ChatOCIGenAI(
    model_id="meta.llama-3.3-70b-instruct",
    service_endpoint="https://inference.generativeai.us-chicago-1.oci.oraclecloud.com",
    compartment_id="ocid1.compartment.oc1..xxxxx",
)

print(llm.invoke("Sing a ballad of LangChain.").content)

ChatOCIGenAI supports streaming, tool calling, structured output, and multimodal inputs. See the langchain-oci README for on-demand vs. dedicated AI cluster models, authentication options, and OCI Data Science model deployments.

Vector search with Oracle Database

import oracledb
from langchain_oci import OCIGenAIEmbeddings
from langchain_oracledb.vectorstores import OracleVS, DistanceStrategy

conn = oracledb.connect(user="user", password="password", dsn="dsn")

embeddings = OCIGenAIEmbeddings(
    model_id="cohere.embed-v4.0",
    service_endpoint="https://inference.generativeai.us-chicago-1.oci.oraclecloud.com",
    compartment_id="ocid1.compartment.oc1..xxxxx",
)

vector_store = OracleVS(conn, embeddings, "my_docs", DistanceStrategy.COSINE)
vector_store.add_texts(["Oracle AI Vector Search is built into Oracle Database."])

results = vector_store.similarity_search("What is AI Vector Search?", k=1)

Durable agents with LangGraph

from langgraph_oracledb.checkpoint.oracle import OracleSaver

with OracleSaver.from_conn_string("user/password@localhost:1521/FREEPDB1") as checkpointer:
    checkpointer.setup()  # create tables & apply migrations once
    graph = builder.compile(checkpointer=checkpointer)
    graph.invoke({"messages": [...]}, config={"configurable": {"thread_id": "1"}})

🎓 Samples

The samples/ directory is a numbered, hands-on learning path — start at 01 and work up, or jump straight to the topic you need. Every track has its own README and runnable code; these same samples are featured on the official LangChain OCI integration page.

Sample	Level	Topics
01: Getting Started	Beginner	Authentication, basic chat, providers
02: Vision & Multimodal	Beginner	Image analysis, PDF, video, audio
03: Building AI Agents	Intermediate	ReAct agents, tools, memory
04: Tool Calling Mastery	Intermediate	Parallel tools, workflows
05: Structured Output	Intermediate	Pydantic schemas, JSON modes
07: Async for Production	Advanced	ainvoke, astream, FastAPI
09: Provider Deep Dive	Specialized	Meta, Gemini, Cohere, xAI specifics
10: Embeddings	Specialized	Text & image embeddings, RAG
11: Deepagents	Specialized	Deep agents with Autonomous Database & OpenSearch datastores

📖 Documentation

Official LangChain integration pages

• OCI provider page — installation, authentication, chat models, embeddings, model deployments, and the samples above
• ChatOCIGenAI and OCIGenAIEmbeddings component guides
• Oracle AI Vector Search provider page — document loaders, text splitter, embeddings, summaries, and the vector store
• Oracle AI Vector Search end-to-end RAG demo

In this repository

• libs/oci/README.md — full langchain-oci guide: on-demand and dedicated AI cluster models, multimodal content, structured output, tool calling, deepagents with datastores, and OCI Data Science deployments
• libs/oracledb/README.md — full langchain-oracledb guide: connecting with python-oracledb, OracleVS with chunking and indexing, loaders, splitter, embeddings, and summaries
• libs/langgraph-oracledb/README.md — checkpointer and store quickstarts (sync and async), plus vector-search store configuration
• libs/js/langchain-oracledb/README.md — LangChain.js usage for every component, with build and test instructions
• samples/README.md — the learning path index, feature coverage matrix, and prerequisites

💁 Contributing

This project welcomes contributions from the community. Contributors must sign the Oracle Contributor Agreement and sign off each commit (git commit -s). See the contribution guide for the full workflow.

🔐 Security

Please report security vulnerabilities privately to Oracle following the security guide — not through public GitHub issues.

📕 License

Copyright (c) 2025 Oracle and/or its affiliates.

Released under the Universal Permissive License v1.0. See LICENSE.txt and THIRD_PARTY_LICENSES.txt for details.