Labelbox Python SDK

7.6.0 · active · verified Wed Apr 15

The Labelbox Python SDK (current version 7.6.0) provides a programmatic interface to interact with the Labelbox platform, enabling users to manage data, projects, labels, and model integrations. It allows for automating data import, export of labeled data, managing labeling projects, and integrating machine learning workflows. The library has a frequent release cadence, with multiple minor versions and bug fixes released throughout the year.

Warnings

gotcha Always store your Labelbox API key securely, preferably as an environment variable (e.g., `LABELBOX_API_KEY`) rather than hardcoding it directly in your code. The `Client` constructor expects the key.
Fix: Use `os.environ.get('LABELBOX_API_KEY')` when initializing `labelbox.Client`.
gotcha When uploading data, ensure your `DataRow` metadata and assets conform to the expected Labelbox formats. Incorrect formatting, especially for complex data types or external URLs, is a common source of upload errors.
Fix: Refer to the official Labelbox documentation for the exact data row and asset payload specifications, including schema requirements for different data types (e.g., images, video, text, geospatial).
breaking While v7.0.0 primarily introduced new features, any significant changes to project ontologies or workflow management (added in v7.x releases) can impact existing labeling pipelines or data exports. Always test with a staging environment.
Fix: Before upgrading to a new major or minor version, review the release notes carefully, especially regarding ontology changes or new workflow features, and test your integration thoroughly in a non-production environment.
gotcha Be mindful of API rate limits when performing bulk operations. While the SDK handles some retries, extensive, rapid requests can lead to `TooManyRequests` errors. Implement exponential backoff for large-scale automation.
Fix: Introduce delays or leverage batch processing features where available for large data operations. Consult Labelbox API documentation for specific rate limit details.

Install

pip install labelbox Install latest version

Imports

Client
```
from labelbox import Client
```
The primary entry point for interacting with the Labelbox API.
Project
```
from labelbox import Project
```
Dataset
```
from labelbox import Dataset
```
DataRow
```
from labelbox import DataRow
```

Quickstart

This quickstart demonstrates how to initialize the Labelbox Client using an API key stored as an environment variable and then retrieve a list of your projects. Ensure `LABELBOX_API_KEY` is set before running.

import os
from labelbox import Client, Project

# Authenticate with your Labelbox API key
# It is recommended to store your API key as an environment variable
# e.g., export LABELBOX_API_KEY="YOUR_API_KEY"
try:
    client = Client(os.environ.get('LABELBOX_API_KEY'))
    print("Successfully connected to Labelbox.")

    # Example: List your projects
    projects = list(client.get_projects())
    if projects:
        print(f"Found {len(projects)} projects:")
        for project in projects:
            print(f"- {project.name} (ID: {project.uid})")
    else:
        print("No projects found in your workspace.")

except Exception as e:
    print(f"An error occurred: {e}")
    print("Please ensure your LABELBOX_API_KEY environment variable is set correctly.")

view raw JSON →