Skip to main content

​
AI SDK

The AI SDK gives you programmatic access to Collate through two complementary paths: MCP Tools for building custom AI applications with any LLM, and AI Studio Agents for invoking ready-to-use AI assistants from Collate’s AI Studio. Available across Python, TypeScript, Java, and a standalone CLI.
You can find the source code for the AI SDK in the GitHub repository. Contributions are always welcome!

​
Explore the AI SDK

Invoke Agent

Send a message to an AI Studio Agent and get a response, with optional streaming and multi-turn conversation support.

Create Agent

Programmatically create an AI Studio Agent with a persona, abilities, and API access.

Create Persona

Create and manage personas that define the behavior and personality of AI Studio Agents.

MCP Tools

Use Collate’s MCP tools to give any LLM access to your metadata through search, lineage, glossary, and entity operations.

​
Available SDKs

SDKPackageInstall
Pythondata-ai-sdkpip install data-ai-sdk
TypeScript@openmetadata/ai-sdknpm install @openmetadata/ai-sdk
Javaorg.open-metadata:ai-sdkMaven / Gradle
CLIai-sdkInstall script

​
Prerequisites

You need:
  1. A Collate instance with AI Studio Agents enabled
  2. A Bot JWT token for API authentication
To get a JWT token, go to Settings > Bots in your Collate instance, select your bot, and copy the token. See How to get the JWT Token for detailed instructions.

​
Configuration

Set the following environment variables: 
export AI_SDK_HOST="https://your-org.getcollate.io"
export AI_SDK_TOKEN="your-bot-jwt-token"
All environment variables:
VariableRequiredDefaultDescription
AI_SDK_HOSTYes-Your Collate server URL
AI_SDK_TOKENYes-Bot JWT token
AI_SDK_TIMEOUTNo120Request timeout in seconds
AI_SDK_VERIFY_SSLNotrueVerify SSL certificates
AI_SDK_MAX_RETRIESNo3Number of retry attempts
AI_SDK_RETRY_DELAYNo1.0Base delay between retries (seconds)

​
Quick Start

​
Python

pip install data-ai-sdk

from ai_sdk import AISdk, AISdkConfig

config = AISdkConfig.from_env()  # reads AI_SDK_HOST and AI_SDK_TOKEN
client = AISdk.from_config(config)

# Invoke an agent
response = client.agent("DataQualityPlannerAgent").call(
    "What data quality tests should I add for the customers table?"
)
print(response.response)

# Stream responses in real time
for event in client.agent("DataQualityPlannerAgent").stream("Analyze the orders table"):
    if event.type == "content":
        print(event.content, end="", flush=True)

​
TypeScript

npm install @openmetadata/ai-sdk

import { AISdk } from '@openmetadata/ai-sdk';

const client = new AISdk({
  host: 'https://your-org.getcollate.io',
  token: 'your-bot-jwt-token'
});

const response = await client.agent('DataQualityPlannerAgent').call(
  'What data quality tests should I add for the customers table?'
);
console.log(response.response);

// Stream responses
for await (const event of client.agent('DataQualityPlannerAgent').stream('Analyze data quality')) {
  if (event.type === 'content') {
    process.stdout.write(event.content || '');
  }
}
Zero runtime dependencies. Works in Node.js 18+, browsers, Deno, and Bun.

​
Java

<dependency>
  <groupId>org.open-metadata</groupId>
  <artifactId>ai-sdk</artifactId>
  <version>0.1.2</version>
</dependency>

import io.openmetadata.ai.AISdk;

AISdk client = new AISdk.Builder()
    .host("https://your-org.getcollate.io")
    .token("your-bot-jwt-token")
    .build();

InvokeResponse response = client.agent("DataQualityPlannerAgent")
    .call("What data quality tests should I add?");
System.out.println(response.getResponse());

​
CLI

# Install
curl -sSL https://raw.githubusercontent.com/open-metadata/ai-sdk/main/cli/install.sh | sh

# Configure
ai-sdk configure

# Invoke an agent
ai-sdk invoke DataQualityPlannerAgent "Analyze the customers table"
The CLI provides an interactive TUI with markdown rendering and syntax highlighting.