Getting started

📢 Note: The API is currently in Beta.

Vectorize provides an API for working with pipelines. Using the API, you can create, delete, schedule, start, and stop pipelines. You can also get details on pipeline events and metrics. The API provides an alternative retrieval endpoint for querying the vector index associated with your pipeline.

With the API, you can also create, update, and delete source connectors, API platforms, and vector DBs programmatically. If you create a file upload connnector—using the API or in the UI—you can also upload files to Vectorize to be used in your pipelines.

For authentication, you must create an API access token in the Vectorize UI. For details, see Managing Access Tokens for Using the API and Retrieval Endpoint.

Notice: The Vectorize API is currently in Beta. Although we will attempt to prevent them, we may need to make breaking changes.

This Getting Started shows you how to install your client of choice and then make your first Vectorize API request.

Install the Vectorize API library

Python

pip install vectorize-client --upgrade

Node.js

npm install @vectorize-io/vectorize-client

Make your first request

1. Claim your access token.

2. Go to the platform to retrieve your organization ID. From the browser url extract the UUID after the organization/ path.

(e.g. for https://platform.vectorize.io/organization/ecf3fa1d-30d0-4df1-8af6-f4852bc851cb the organization ID is ecf3fa1d-30d0-4df1-8af6-f4852bc851cb).

1. Use the following code to list your pipelines:

Python

import vectorize_client as v

org="your-org-id"
token="your-token"

api = v.ApiClient(v.Configuration(access_token=token))
pipelines = v.PipelinesApi(api)

response = pipelines.get_pipelines(org)
for pipeline in response.data:
    print("Pipeline: " + pipeline.name)

Node.js

const {Configuration, PipelinesApi, ExtractionApi, UploadsApi, ConnectorsApi, FilesApi, SourceConnectorType, DestinationConnectorType, AIPlatformType} = require("@vectorize-io/vectorize-client");
const fs = require("fs");

const token = "your-token"
const org = "your-org-id"

async function main() {
    try {
        const api = new Configuration({
            accessToken: token
        });
        const pipelinesApi = new PipelinesApi(api);
        const uploadsApi = new UploadsApi(api);
        const connectorsApi = new ConnectorsApi(api);
        const extractionApi = new ExtractionApi(api);
        const filesApi = new FilesApi(api);

        const response = await pipelinesApi.getPipelines({
            organization: org
        });

        for (const pipeline of response.data) {
            console.log("Pipeline: " + pipeline.name)
        }
    } catch (error) {
        console.error(error?.response);
        console.error(await error?.response?.text());
        throw error;
    }
}

main();

Next steps

• Create a RAG Pipeline, perform Vector Search and more
• Perform text extraction from unstructured Data using Vectorize Iris
• Connect AI assistants using the MCP Server

Looking for a full working Python example? You can open it on Google Colab or download it from the GitHub repository

Looking for a full working Node.js example? You can find it on the GitHub repository

A complete reference for the Vectorize API can be found here: Vectorize API Reference. This site includes tools for interactively testing API calls as well as generating code in various programming languages.