---
id: supervisely
version: "6.74.28"
license: unclear
license_treatment: unclear
maintenance: active
---

# supervisely — Supervisely Python SDK.

License: unclear · Maintenance: active · Downloads: 687.6K/mo

## What it is and what it does

Supervisely is a Python SDK that wraps the Supervisely platform's HTTP REST API, enabling programmatic control over computer vision workflows including data labeling, project management, model deployment, and app creation. The platform itself is a web-based operating system for computer vision tasks—supporting image, video, 3D point cloud, and medical image (DICOM) annotation, model training, and collaborative workflows. The SDK lets developers automate routine tasks, integrate custom models, manage datasets, and build applications that run within the Supervisely ecosystem.

The package is production-stable and actively maintained, with broad Python version support. Its dependency tree includes standard CV libraries (numpy, opencv-python, pillow, SimpleITK, pydicom, trimesh) and web infrastructure (fastapi, starlette, uvicorn, websockets), reflecting its role as both a client library and a foundation for building interactive web-based apps. Authentication is token-based and typically environment-driven.

Use it for:

- Automate bulk upload and annotation of images or videos to Supervisely projects from local or cloud storage.
- Build headless scripts that train models, evaluate performance, and push predictions back to the platform for review.
- Create custom web apps with interactive UIs that integrate into Supervisely's labeling tools and ecosystem.
- Download labeled datasets and annotations in standard formats for local model development and experimentation.
- Manage team collaboration by programmatically creating projects, datasets, and assigning tasks to team members.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Python SDK for the Supervisely computer vision platform, providing programmatic access to data labeling, model management, project workflows, and app deployment capabilities.

Yes, if you are working within the Supervisely ecosystem or planning to integrate computer vision workflows with it. The SDK is actively maintained, has low install friction, and provides a mature Python interface to a production platform. However, clarify the license implications beforehand (treatment is currently unclear), and verify that the 53 dependencies align with your environment constraints. Not relevant if you need a standalone CV library independent of the Supervisely platform.

## Install

pip install supervisely
uv add supervisely
poetry add supervisely

## Installing supervisely

Before you install: Low install friction with a pure Python wheel. Actively maintained with a release 2 days ago and 540 repository stars. Supports modern Python versions (3.8–3.14) and carries 53 runtime dependencies including core CV libraries (numpy, opencv-python, pillow) and web frameworks (fastapi, starlette).

Quickstart:

pip install supervisely

import supervisely as sly
api = sly.Api.from_env()
project = api.project.create(workspace_id=123, name="demo project")

Requires a Supervisely account and valid API token (typically set via environment variable for Api.from_env()).

Verify before relying:

- Whether the package's license treatment ('unclear') poses any legal constraint for commercial use.
- Performance characteristics and scalability limits for large-scale data uploads or model inference.
- Whether all 53 runtime dependencies are required for basic SDK usage or if many are optional.

## Package facts

- License: not declared (unclear)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 687.6K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

computer vision platform sdk, image annotation api python, data labeling automation, model deployment framework, supervisely integration, computer-vision, data-labeling, platform-sdk

[View on SkillFed](https://skillfed.io/packages/supervisely) · [View on PyPI](https://pypi.org/project/supervisely/)