---
id: unstructured-client
version: "0.46.1"
license: MIT
license_treatment: permissive
maintenance: active
---

# unstructured-client — Python Client SDK for Unstructured API

License: permissive · Maintenance: active · Downloads: 8.9M/mo

## What it is and what it does

unstructured-client is a Python HTTP client for the Unstructured Platform API, a cloud service that extracts structured data from unstructured documents. It wraps the Platform's Partition and Workflow endpoints, allowing you to send documents to the API and retrieve parsed, structured output. The SDK handles authentication, retries, error handling, and file uploads.

You use it by instantiating an UnstructuredClient, then calling methods on its resource objects (e.g., destinations, partitions) to submit documents and retrieve results. The SDK depends on httpx for HTTP transport, pydantic for data validation, and pypdf/pypdfium2 for local PDF handling. It supports custom retry strategies and detailed error reporting via UnstructuredClientError and its subclasses.

Use it for:

- Extract structured data from PDFs, images, or other documents by sending them to the Unstructured Platform API.
- Build document processing pipelines that parse invoices, contracts, or forms and return JSON-structured output.
- Integrate document parsing into a larger ETL workflow using the Workflow Endpoint.
- Handle file uploads and manage API responses with built-in retry logic and error handling.
- Prototype document extraction without maintaining local parsing infrastructure.

## Worth the install?

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

HTTP client SDK for the Unstructured Platform API that processes documents and extracts structured data from unstructured content.

Yes, if you need to extract structured data from unstructured documents and are willing to use a cloud API. The SDK is actively maintained, has low install friction, carries no known vulnerabilities, and uses a permissive license. Requires Python 3.11+. Verify that the free tier and API pricing fit your volume and use case before committing.

## Install

pip install unstructured-client
uv add unstructured-client
poetry add unstructured-client

## Installing unstructured-client

Before you install: Low friction installation with a pure-Python wheel. Active maintenance with a release 9 days ago. Requires Python 3.11 or later. Seven runtime dependencies are all standard HTTP and data-handling libraries (httpx, pydantic, pypdf, pypdfium2, requests-toolbelt, aiofiles, httpcore).

License in practice: MIT license is permissive; you can use this package freely in commercial and private projects with minimal restrictions.

Quickstart:

pip install unstructured-client

from unstructured_client import UnstructuredClient

with UnstructuredClient() as client:
    res = client.destinations.create_connection_check_destinations(
        request={"destination_id": "cb9e35c1-0b04-4d98-83fa-fa6241323f96"}
    )

Requires Python 3.11 or later. API key or credentials for the Unstructured Platform are needed to authenticate requests.

Verify before relying:

- Whether the free tier (1000 pages per day for 14 days) is sufficient for typical use cases.
- Performance characteristics when processing large documents or high-volume batches.
- Supported file formats beyond PDF (mentioned in dependencies but not detailed in excerpt).

## Package facts

- License: MIT (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 8.9M/month (top 5,000 on PyPI)
- Known vulnerabilities: none known

## Tags

unstructured data extraction sdk, document processing api client, http client for unstructured platform, extract data from documents, unstructured content parser, api sdk for document parsing, structured data from files, document-extraction, api-client, data-parsing

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