langchain-hatidata

LangChain integration for HatiData -- Memory, VectorStore, and Tools backed by HatiData's in-VPC data warehouse with agent-aware billing and audit.

Installation

pip install langchain-hatidata

Quick Start

from langchain_hatidata import HatiDataMemory, HatiDataVectorStore, HatiDataToolkit

# Conversation memory persisted to HatiData
memory = HatiDataMemory(host="proxy.internal", agent_id="my-agent")

# Vector store for RAG
vectorstore = HatiDataVectorStore(
    host="proxy.internal",
    agent_id="my-agent",
    table="embeddings",
)

# SQL tools for agents
toolkit = HatiDataToolkit(host="proxy.internal", agent_id="my-agent")

Features

• HatiDataMemory -- Persistent conversation memory backed by HatiData, with full audit trails
• HatiDataVectorStore -- Vector similarity search for RAG pipelines, stored in-VPC
• HatiDataToolkit -- SQL query, table listing, schema inspection, and context search tools for LangChain agents
• Agent-aware billing -- Every query is tagged with agent ID and framework for per-agent cost tracking
• Sub-10ms latency -- Queries execute in-VPC with no data leaving your network

Components

Documentation

Full documentation is available at docs.hatiosai.com/hatidata.

License

Apache License 2.0. Copyright (c) Marviy Pte Ltd. See LICENSE for details.