By AI Surface · LlamaIndex

LlamaIndex + Connect AI

Build ReAct agents that reason across live business data. LlamaIndex wraps Connect AI’s MCP tools as native functions, works with OpenAI and Anthropic out of the box.

LlamaIndex getting started 

from llama_index.core.agent.workflow import ReActAgent
from llama_index.core.tools import FunctionTool

# Wrap Connect AI MCP tools for the ReAct agent
tools = [FunctionTool.from_defaults(fn=t)
         for t in mcp_client.list_tools()]

agent = ReActAgent(tools=tools, llm=llm)

response = await agent.run(
    "Which accounts had the most support "
    "tickets last quarter?"
)
print(response)

Developer guide

Full step-by-step guide to building a production-ready LlamaIndex ReAct agent with Connect AI MCP.

LlamaIndex ReAct Agent: Getting Started Build AI agents with LlamaIndex ReAct framework using Connect AI as the SQL data tool. Includes full code walkthrough and runnable examples.

Sources for LlamaIndex

Common LlamaIndex + source patterns. Tap any combination to see the tutorial.

Salesforce CRM pipeline research and account briefings with ReAct multi-step reasoning.
Workday HR data queries, org analysis, and headcount insights via tool orchestration.
HubSpot CRM research and deal intelligence via tool-based reasoning.
NetSuite ERP queries for financial reporting and inventory analysis.
Snowflake Warehouse analytics and KPI questions via tool-based reasoning.
Postgres Operational database research and schema-aware querying.

Sample agent

A polished LlamaIndex ReAct reference. Looking for a quick start in a different framework? See the npx scaffold (LangChain, Google ADK, or Vercel AI SDK).

CDataSoftware/connectai-llamaindex-agent

Production-ready LlamaIndex ReAct agent with Connect AI MCP, multi-turn context, and support for OpenAI and Anthropic LLMs.

View on GitHub

Evaluating Connect AI for your team? Learn more about Connect AI.

Connect AI documentation

Full API reference, authentication guides, and configuration details.

View docs