How to integrate Scale ai MCP with LlamaIndex

GET STARTED FOR FREE
Framework Integration Gradient
Scale ai Logo
LlamaIndex Logo
divider

Introduction

This guide walks you through connecting Scale ai to LlamaIndex using the Composio tool router. By the end, you'll have a working Scale ai agent that can create image labeling task for dataset 'road-signs', list completed annotation tasks for project, fetch results of data labeling job through natural language commands.

This guide will help you understand how to give your LlamaIndex agent real control over a Scale ai account through Composio's Scale ai MCP server.

Before we dive in, let's take a quick look at the key ideas and tools involved.

TL;DR

Here's what you'll learn:
  • Set your OpenAI and Composio API keys
  • Install LlamaIndex and Composio packages
  • Create a Composio Tool Router session for Scale ai
  • Connect LlamaIndex to the Scale ai MCP server
  • Build a Scale ai-powered agent using LlamaIndex
  • Interact with Scale ai through natural language

What is LlamaIndex?

LlamaIndex is a data framework for building LLM applications. It provides tools for connecting LLMs to external data sources and services through agents and tools.

Key features include:

  • ReAct Agent: Reasoning and acting pattern for tool-using agents
  • MCP Tools: Native support for Model Context Protocol
  • Context Management: Maintain conversation context across interactions
  • Async Support: Built for async/await patterns

What is the Scale ai MCP server, and what's possible with it?

The Scale ai MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to your Scale ai account. It provides structured and secure access so your agent can perform Scale ai operations on your behalf.

Supported Tools & Triggers

Tools
Add Studio AssignmentsTool to add project assignments to team members in Scale AI Studio.
Add Task TagsTool to add tags to an existing task.
Create BatchTool to create a new batch within a project.
Create Document Transcription TaskTool to create a document transcription task where workers transcribe and annotate information from single or multi-page documents.
Create Image Annotation TaskTool to create an image annotation task where annotators label images with vector geometric shapes (box, polygon, line, point, cuboid, ellipse).
Create Lidar Annotation TaskTool to create a lidar annotation task where annotators mark objects with 3D cuboids in 3D space.
Create LiDAR Segmentation TaskTool to create a LiDAR segmentation task where annotators assign semantic class labels to individual LiDAR points.
Create Named Entity Recognition TaskTool to create a named entity recognition task for labelers to highlight text entity mentions.
Create Segmentation Annotation TaskTool to create a segmentation task where annotators classify pixels in an image according to provided labels.
Create Text Collection TaskTool to create a textcollection task for collecting information from attachments and/or web sources.
Create Video Annotation TaskTool to create a video annotation task where annotators draw geometric shapes around specified objects across video frames.
Create Video Playback Annotation TaskTool to create a video playback annotation task where annotators draw shapes around specified objects in video files.
Delete Task TagsTool to remove specified tags from a Scale AI task.
Delete Task Unique IDTool to remove the unique identifier from a task.
Finalize BatchTool to finalize a batch so its tasks can be worked on.
Get AssetsTool to retrieve file assets with filtering capabilities by project and metadata.
Get BatchTool to retrieve the details of a batch with the specified name.
Get Batch StatusTool to retrieve the current status of a batch and task completion counts.
Get Fixless AuditsTool to retrieve fixless audits by task ID or audit ID.
Get ProjectTool to retrieve details about a specific Scale AI project using its unique identifier.
Get Quality LabelersTool to retrieve training attempts matching provided filter parameters.
Get Studio AssignmentsTool to retrieve current project assignments of all active team users in Scale AI Studio.
Get Studio BatchesTool to retrieve basic information about all pending batches in Studio.
Get TaskTool to retrieve detailed information about a specific task in Scale AI.
Get TeamsTool to retrieve basic information about all team members associated with the account.
Get Task by IDTool to retrieve detailed information about a specific task using its task ID.
Get Secure Task Response URLTool to retrieve secure authenticated task response data.
Import FileTool to import files from an external URL endpoint into Scale's system rather than uploading directly from local storage.
Invite Team MemberTool to invite users by email to team with specified role.
List BatchesTool to retrieve all batches in descending order by creation date.
List ProjectsTool to retrieve information for all projects in the Scale AI account with optional archived filtering.
List TasksTool to retrieve a paginated list of tasks in descending order by creation time.
Re-send Task CallbackTool to re-send a callback for a completed or errored task to the callback_url.
Remove Studio AssignmentsTool to unassign projects from specified team members in Scale AI Studio.
Reset Batch PrioritiesTool to restore batch priority order to default order (calibration batches first, then sorted by creation date).
Set Batch PrioritiesTool to modify batch priority order in Scale AI Studio.
Set Project OntologyTool to set ontologies on a Scale AI project.
Set Project ParametersTool to set default parameters for tasks created under a project.
Set Task MetadataTool to set key-value metadata on an existing Scale AI task.
Update Task Unique IDTool to update or assign a unique identifier to a task.
Upload FileTool to upload a local file to Scale's servers with a maximum size limit of 80 MB per file.

What is the Composio tool router, and how does it fit here?

What is Tool Router?

Composio's Tool Router helps agents find the right tools for a task at runtime. You can plug in multiple toolkits (like Gmail, HubSpot, and GitHub), and the agent will identify the relevant app and action to complete multi-step workflows. This can reduce token usage and improve the reliability of tool calls. Read more here: Getting started with Tool Router

The tool router generates a secure MCP URL that your agents can access to perform actions.

How the Tool Router works

The Tool Router follows a three-phase workflow:

  1. Discovery: Searches for tools matching your task and returns relevant toolkits with their details.
  2. Authentication: Checks for active connections. If missing, creates an auth config and returns a connection URL via Auth Link.
  3. Execution: Executes the action using the authenticated connection.

Step-by-step Guide

Prerequisites

Before you begin, make sure you have:
  • Python 3.8/Node 16 or higher installed
  • A Composio account with the API key
  • An OpenAI API key
  • A Scale ai account and project
  • Basic familiarity with async Python/Typescript

Getting API Keys for OpenAI, Composio, and Scale ai

OpenAI API key (OPENAI_API_KEY)
  • Go to the OpenAI dashboard
  • Create an API key if you don't have one
  • Assign it to OPENAI_API_KEY in .env
Composio API key and user ID
  • Log into the Composio dashboard
  • Copy your API key from Settings
    • Use this as COMPOSIO_API_KEY
  • Pick a stable user identifier (email or ID)
    • Use this as COMPOSIO_USER_ID

Installing dependencies

pip install composio-llamaindex llama-index llama-index-llms-openai llama-index-tools-mcp python-dotenv

Create a new Python project and install the necessary dependencies:

  • composio-llamaindex: Composio's LlamaIndex integration
  • llama-index: Core LlamaIndex framework
  • llama-index-llms-openai: OpenAI LLM integration
  • llama-index-tools-mcp: MCP client for LlamaIndex
  • python-dotenv: Environment variable management

Set environment variables

bash
OPENAI_API_KEY=your-openai-api-key
COMPOSIO_API_KEY=your-composio-api-key
COMPOSIO_USER_ID=your-user-id

Create a .env file in your project root:

These credentials will be used to:

  • Authenticate with OpenAI's GPT-5 model
  • Connect to Composio's Tool Router
  • Identify your Composio user session for Scale ai access

Import modules

import asyncio
import os
import dotenv

from composio import Composio
from composio_llamaindex import LlamaIndexProvider
from llama_index.core.agent.workflow import ReActAgent
from llama_index.core.workflow import Context
from llama_index.llms.openai import OpenAI
from llama_index.tools.mcp import BasicMCPClient, McpToolSpec

dotenv.load_dotenv()

Create a new file called scale ai_llamaindex_agent.py and import the required modules:

Key imports:

  • asyncio: For async/await support
  • Composio: Main client for Composio services
  • LlamaIndexProvider: Adapts Composio tools for LlamaIndex
  • ReActAgent: LlamaIndex's reasoning and action agent
  • BasicMCPClient: Connects to MCP endpoints
  • McpToolSpec: Converts MCP tools to LlamaIndex format

Load environment variables and initialize Composio

OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
COMPOSIO_API_KEY = os.getenv("COMPOSIO_API_KEY")
COMPOSIO_USER_ID = os.getenv("COMPOSIO_USER_ID")

if not OPENAI_API_KEY:
    raise ValueError("OPENAI_API_KEY is not set in the environment")
if not COMPOSIO_API_KEY:
    raise ValueError("COMPOSIO_API_KEY is not set in the environment")
if not COMPOSIO_USER_ID:
    raise ValueError("COMPOSIO_USER_ID is not set in the environment")

What's happening:

This ensures missing credentials cause early, clear errors before the agent attempts to initialise.

Create a Tool Router session and build the agent function

async def build_agent() -> ReActAgent:
    composio_client = Composio(
        api_key=COMPOSIO_API_KEY,
        provider=LlamaIndexProvider(),
    )

    session = composio_client.create(
        user_id=COMPOSIO_USER_ID,
        toolkits=["scale_ai"],
    )

    mcp_url = session.mcp.url
    print(f"Composio MCP URL: {mcp_url}")

    mcp_client = BasicMCPClient(mcp_url, headers={"x-api-key": COMPOSIO_API_KEY})
    mcp_tool_spec = McpToolSpec(client=mcp_client)
    tools = await mcp_tool_spec.to_tool_list_async()

    llm = OpenAI(model="gpt-5")

    description = "An agent that uses Composio Tool Router MCP tools to perform Scale ai actions."
    system_prompt = """
    You are a helpful assistant connected to Composio Tool Router.
    Use the available tools to answer user queries and perform Scale ai actions.
    """
    return ReActAgent(tools=tools, llm=llm, description=description, system_prompt=system_prompt, verbose=True)

What's happening here:

  • We create a Composio client using your API key and configure it with the LlamaIndex provider
  • We then create a tool router MCP session for your user, specifying the toolkits we want to use (in this case, scale ai)
  • The session returns an MCP HTTP endpoint URL that acts as a gateway to all your configured tools
  • LlamaIndex will connect to this endpoint to dynamically discover and use the available Scale ai tools.
  • The MCP tools are mapped to LlamaIndex-compatible tools and plug them into the Agent.

Create an interactive chat loop

async def chat_loop(agent: ReActAgent) -> None:
    ctx = Context(agent)
    print("Type 'quit', 'exit', or Ctrl+C to stop.")

    while True:
        try:
            user_input = input("\nYou: ").strip()
        except (KeyboardInterrupt, EOFError):
            print("\nBye!")
            break

        if not user_input or user_input.lower() in {"quit", "exit"}:
            print("Bye!")
            break

        try:
            print("Agent: ", end="", flush=True)
            handler = agent.run(user_input, ctx=ctx)

            async for event in handler.stream_events():
                # Stream token-by-token from LLM responses
                if hasattr(event, "delta") and event.delta:
                    print(event.delta, end="", flush=True)
                # Show tool calls as they happen
                elif hasattr(event, "tool_name"):
                    print(f"\n[Using tool: {event.tool_name}]", flush=True)

            # Get final response
            response = await handler
            print()  # Newline after streaming
        except KeyboardInterrupt:
            print("\n[Interrupted]")
            continue
        except Exception as e:
            print(f"\nError: {e}")

What's happening here:

  • We're creating a direct terminal interface to chat with your Scale ai database
  • The LLM's responses are streamed to the CLI for faster interaction.
  • The agent uses context to maintain conversation history
  • You can type 'quit' or 'exit' to stop the chat loop gracefully
  • Agent responses and any errors are displayed in a clear, readable format

Define the main entry point

async def main() -> None:
    agent = await build_agent()
    await chat_loop(agent)

if __name__ == "__main__":
    # Handle Ctrl+C gracefully
    signal.signal(signal.SIGINT, lambda s, f: (print("\nBye!"), exit(0)))
    try:
        asyncio.run(main())
    except KeyboardInterrupt:
        print("\nBye!")

What's happening here:

  • We're orchestrating the entire application flow
  • The agent gets built with proper error handling
  • Then we kick off the interactive chat loop so you can start talking to Scale ai

Run the agent

npx ts-node llamaindex-agent.ts

When prompted, authenticate and authorise your agent with Scale ai, then start asking questions.

Complete Code

Here's the complete code to get you started with Scale ai and LlamaIndex:

import asyncio
import os
import signal
import dotenv

from composio import Composio
from composio_llamaindex import LlamaIndexProvider
from llama_index.core.agent.workflow import ReActAgent
from llama_index.core.workflow import Context
from llama_index.llms.openai import OpenAI
from llama_index.tools.mcp import BasicMCPClient, McpToolSpec

dotenv.load_dotenv()

OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
COMPOSIO_API_KEY = os.getenv("COMPOSIO_API_KEY")
COMPOSIO_USER_ID = os.getenv("COMPOSIO_USER_ID")

if not OPENAI_API_KEY:
    raise ValueError("OPENAI_API_KEY is not set")
if not COMPOSIO_API_KEY:
    raise ValueError("COMPOSIO_API_KEY is not set")
if not COMPOSIO_USER_ID:
    raise ValueError("COMPOSIO_USER_ID is not set")

async def build_agent() -> ReActAgent:
    composio_client = Composio(
        api_key=COMPOSIO_API_KEY,
        provider=LlamaIndexProvider(),
    )

    session = composio_client.create(
        user_id=COMPOSIO_USER_ID,
        toolkits=["scale_ai"],
    )

    mcp_url = session.mcp.url
    print(f"Composio MCP URL: {mcp_url}")

    mcp_client = BasicMCPClient(mcp_url, headers={"x-api-key": COMPOSIO_API_KEY})
    mcp_tool_spec = McpToolSpec(client=mcp_client)
    tools = await mcp_tool_spec.to_tool_list_async()

    llm = OpenAI(model="gpt-5")
    description = "An agent that uses Composio Tool Router MCP tools to perform Scale ai actions."
    system_prompt = """
    You are a helpful assistant connected to Composio Tool Router.
    Use the available tools to answer user queries and perform Scale ai actions.
    """
    return ReActAgent(
        tools=tools,
        llm=llm,
        description=description,
        system_prompt=system_prompt,
        verbose=True,
    );

async def chat_loop(agent: ReActAgent) -> None:
    ctx = Context(agent)
    print("Type 'quit', 'exit', or Ctrl+C to stop.")

    while True:
        try:
            user_input = input("\nYou: ").strip()
        except (KeyboardInterrupt, EOFError):
            print("\nBye!")
            break

        if not user_input or user_input.lower() in {"quit", "exit"}:
            print("Bye!")
            break

        try:
            print("Agent: ", end="", flush=True)
            handler = agent.run(user_input, ctx=ctx)

            async for event in handler.stream_events():
                # Stream token-by-token from LLM responses
                if hasattr(event, "delta") and event.delta:
                    print(event.delta, end="", flush=True)
                # Show tool calls as they happen
                elif hasattr(event, "tool_name"):
                    print(f"\n[Using tool: {event.tool_name}]", flush=True)

            # Get final response
            response = await handler
            print()  # Newline after streaming
        except KeyboardInterrupt:
            print("\n[Interrupted]")
            continue
        except Exception as e:
            print(f"\nError: {e}")

async def main() -> None:
    agent = await build_agent()
    await chat_loop(agent)

if __name__ == "__main__":
    # Handle Ctrl+C gracefully
    signal.signal(signal.SIGINT, lambda s, f: (print("\nBye!"), exit(0)))
    try:
        asyncio.run(main())
    except KeyboardInterrupt:
        print("\nBye!")

Conclusion

You've successfully connected Scale ai to LlamaIndex through Composio's Tool Router MCP layer. Key takeaways:
  • Tool Router dynamically exposes Scale ai tools through an MCP endpoint
  • LlamaIndex's ReActAgent handles reasoning and orchestration; Composio handles integrations
  • The agent becomes more capable without increasing prompt size
  • Async Python provides clean, efficient execution of agent workflows
You can easily extend this to other toolkits like Gmail, Notion, Stripe, GitHub, and more by adding them to the toolkits parameter.

How to build Scale ai MCP Agent with another framework

OpenAI Agents SDK

OpenAI Agents SDK

Use Scale ai MCP with OpenAI Agents SDK

Claude Agent SDK

Claude Agent SDK

Use Scale ai MCP with Claude Agent SDK

Google ADK

Google ADK

Use Scale ai MCP with Google ADK

LangChain

LangChain

Use Scale ai MCP with LangChain

Vercel AI SDK

Vercel AI SDK

Use Scale ai MCP with Vercel AI SDK

Mastra AI

Mastra AI

Use Scale ai MCP with Mastra AI

CrewAI

CrewAI

Use Scale ai MCP with CrewAI

Claude Code

Claude Code

Use Scale ai MCP with Claude Code

Codex

Codex

Use Scale ai MCP with Codex

FAQ

What are the differences in Tool Router MCP and Scale ai MCP?

With a standalone Scale ai MCP server, the agents and LLMs can only access a fixed set of Scale ai tools tied to that server. However, with the Composio Tool Router, agents can dynamically load tools from Scale ai and many other apps based on the task at hand, all through a single MCP endpoint.

Can I use Tool Router MCP with LlamaIndex?

Yes, you can. LlamaIndex fully supports MCP integration. You get structured tool calling, message history handling, and model orchestration while Tool Router takes care of discovering and serving the right Scale ai tools.

Can I manage the permissions and scopes for Scale ai while using Tool Router?

Yes, absolutely. You can configure which Scale ai scopes and actions are allowed when connecting your account to Composio. You can also bring your own OAuth credentials or API configuration so you keep full control over what the agent can do.

How safe is my data with Composio Tool Router?

All sensitive data such as tokens, keys, and configuration is fully encrypted at rest and in transit. Composio is SOC 2 Type 2 compliant and follows strict security practices so your Scale ai data and credentials are handled as safely as possible.

Used by agents from

Context
Letta
glean
HubSpot
Agent.ai
Altera
DataStax
Entelligence
Rolai
Context
Letta
glean
HubSpot
Agent.ai
Altera
DataStax
Entelligence
Rolai
Context
Letta
glean
HubSpot
Agent.ai
Altera
DataStax
Entelligence
Rolai

Never worry about agent reliability

We handle tool reliability, observability, and security so you never have to second-guess an agent action.

Book a call↗Get started for free