Quick Start

Start building with the Arivu SDK.

Arivu SDK Quick Start

Get started with the Arivu SDK to build AI-powered agents connected to your databases. The SDK natively integrates with LLMs, memory, and your SQL environment.

Installation

1. Install Arivu

   Use pip to install the core arivu package.

   pip

   pip install arivu
   
2. Connect to your Database

   Initialize the Arivu client with your database credentials.

   python

   from arivu import Arivu
   
   db = Arivu.connect(
       host="localhost",
       port=5432,
       user="postgres",
       password="password",
       dbname="mydb",
       mode="user",            # 'user' mode limits destructive queries
       dialect="postgresql",
   )
   
3. Run the Pipeline

   Pass a natural language string into the pipeline query interface to securely interact with the database.

   python

   from arivu.pipeline import run_pipeline
   
   pipeline_input = db.query("Show me the top 10 customers by total spend")
   result = run_pipeline(pipeline_input)
   
   if result.success:
       print(result.response)
       print(f"Generated SQL: {result.sql}")
   else:
       print(result.error)
   

Architecture Overview

Arivu translates natural language intents into secure, constrained SQL operations. The pipeline incorporates RLHF verification layers, dynamic schema lookup, and rich memory integration.

Databases

Connect seamlessly to PostgeSQL, SQLite, and MySQL.

LLM Configuration

Configure provider settings and adapt the reasoning engines.

Application Integrations

Expose your agent through REST API, Slack, Telegram, WhatsApp, or Discord integrations.