parallel-web

What it is and what it does

parallel-web is an official Python SDK for the Parallel REST API, generated by Stainless. It offers both synchronous (Parallel) and asynchronous (AsyncParallel) clients powered by httpx, with full type definitions for all request parameters and response fields using Pydantic models and TypedDicts. The library handles authentication via API key, automatic retries for transient errors (2 times by default), configurable timeouts (1 minute default), and detailed error classification (connection errors, rate limits, authentication failures, etc.).

Typical usage involves instantiating a client with an API key, calling methods like client.task_run.create() with typed parameters, and receiving Pydantic response objects with helper methods for serialization. The library supports both blocking and async workflows, optional aiohttp backend for improved concurrent performance, and provides logging via the standard library logging module when PARALLEL_LOG is set to info or debug.

Use it for:

• Build Python applications that call the Parallel API for task processing without manually constructing HTTP requests or parsing responses.
• Integrate Parallel AI capabilities into async/await codebases with the AsyncParallel client and optional aiohttp backend.
• Leverage full IDE autocomplete and type checking for Parallel API parameters and responses using Pydantic models.
• Handle API errors gracefully with specific exception types (RateLimitError, AuthenticationError, etc.) and automatic retry logic.
• Query task results and access response metadata using Pydantic helper methods like model.to_json() and model.to_dict().

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Provides a Python client library for the Parallel REST API with synchronous and asynchronous interfaces, full type hints, and automatic error handling and retries.

Yes. The library is actively maintained (last commit 2 days old), has low install friction, carries a permissive MIT license, requires only common dependencies, and provides a well-structured, type-safe interface to the Parallel API. No known security vulnerabilities. Suitable for production use if your application needs to call the Parallel API.

Install

parallel-web on PyPI

pip install parallel-web

uv add parallel-web

poetry add parallel-web

Installing parallel-web

Before you install

Low friction install with a pure-Python wheel and six common dependencies (anyio, distro, httpx, pydantic, sniffio, typing-extensions). Active maintenance: last commit 2026-08-12, release 2 days old.

License in practice

MIT license (permissive) — you can use, modify, and distribute this library freely in commercial and private projects with minimal restrictions.

Quickstart

pip install parallel-web

import os
from parallel import Parallel

client = Parallel(api_key=os.environ.get("PARALLEL_API_KEY"))
task_run = client.task_run.create(
    input="What was the GDP of France in 2023?",
    processor="base",
)
print(task_run.interaction_id)

Requires Python 3.9 or later; a valid Parallel API key (passed via PARALLEL_API_KEY environment variable or api_key argument).

Verify before relying

• Whether the Parallel API service itself is production-ready and has documented SLAs.
• Whether aiohttp extra (optional dependency for improved async concurrency) is recommended for typical workloads.

Package facts

Evidence: parallel_web-1.3.0-py3-none-any.whl

Tags

More Python Modules packages