aioscraper

High-performance asynchronous Python framework for large-scale API data collection.

API-first. aioscraper orchestrates thousands of concurrent JSON/REST calls with adaptive rate limiting, retries, priority queues and item pipelines. Selectors and a crawling engine are not part of the core - plug in your own parser if you need one (see examples/quotes.py).

Beta notice: APIs and behavior may change; expect sharp edges while things settle.

Table of Contents

• What is aioscraper?
• Key Features
• Installation
• Quick Start
• Examples
• Why aioscraper?
• Use Cases
• Performance
• Documentation
• Changelog
• Contributing

What is aioscraper?

aioscraper is an async Python framework designed for mass data collection from APIs and external services at scale.

Built for:

• Fetching data from hundreds/thousands of REST API endpoints concurrently
• Integrating multiple external services (payment gateways, analytics APIs, etc.)
• Building data aggregation pipelines from heterogeneous API sources
• Queue-based scraping workers consuming tasks from Redis/RabbitMQ
• Microservice fan-out requests with automatic rate limiting and retries

NOT built for:

• Parsing HTML/CSS (but nothing stops you from using BeautifulSoup if you want - see examples/quotes.py)
• Single API requests (use httpx or aiohttp directly)
• GraphQL or WebSocket scraping (different paradigm)

Think: "I need to fetch data from 10,000 product API endpoints" or "I need to poll 50 microservices every minute" → aioscraper is for you.

Key Features

• Async-first core with pluggable HTTP backends (aiohttp/httpx/httpx2) and aiojobs scheduling
• Declarative flow: requests → callbacks → pipelines, with middleware hooks at each stage
• Priority queueing with backpressure, a global concurrency limit and per-group rate limits
• Adaptive rate limiting with EWMA + AIMD algorithm - automatically backs off on server overload
• Small, explicit API that is easy to test and compose with existing async applications

Installation

Choose your HTTP backend:

# Option 1: Use aiohttp (recommended for most cases)
pip install "aioscraper[aiohttp]"

# Option 2: Use httpx (if you prefer httpx ecosystem)
pip install "aioscraper[httpx]"

# Option 3: Use httpx2, the Pydantic-maintained fork of httpx
pip install "aioscraper[httpx2]"

# Option 4: Install several backends for flexibility
pip install "aioscraper[aiohttp,httpx,httpx2]"

Quick Start

Create scraper.py:

import logging
from aioscraper import AIOScraper, Request, Response, ScheduleRequest, Pipeline
from dataclasses import dataclass

logger = logging.getLogger("github_repos")
scraper = AIOScraper()


@dataclass(slots=True)
class RepoStats:
    name: str
    stars: int
    language: str


# registers the pipeline that handles RepoStats items
@scraper.pipeline(RepoStats)
class StatsPipeline:
    def __init__(self):
        self.total_stars = 0

    async def put_item(self, item: RepoStats) -> RepoStats:
        # runs once per extracted item: store it, queue it, validate it, or aggregate as here
        self.total_stars += item.stars
        logger.info("✓ %s: ⭐ %s (%s)", item.name, item.stars, item.language)
        return item

    async def close(self):
        # runs once when the scraper stops: flush buffers, close connections, report totals
        logger.info("Total stars collected: %s", self.total_stars)


# registers an entry point; schedule_request is injected by parameter name
@scraper
async def get_repos(schedule_request: ScheduleRequest):
    repos = (
        "django/django",
        "fastapi/fastapi",
        "pallets/flask",
        "encode/httpx",
        "aio-libs/aiohttp",
    )

    for repo in repos:
        await schedule_request(
            Request(
                url=f"https://api.github.com/repos/{repo}",
                callback=parse_repo,  # runs on a response with a status below 400
                errback=on_failure,  # runs on anything else: 4xx/5xx, timeouts, connection failures
                cb_kwargs={"repo": repo},  # extra arguments for both of them
                headers={"Accept": "application/vnd.github+json"},  # required by the GitHub API
            )
        )


async def parse_repo(response: Response, pipeline: Pipeline):
    # the body has to be read here: the connection is released when the callback returns
    data = await response.json()
    await pipeline(
        RepoStats(
            name=data["full_name"],
            stars=data["stargazers_count"],
            language=data.get("language", "Unknown"),
        )
    )


async def on_failure(exc: Exception, repo: str):
    logger.error("%s: cannot parse response: %s", repo, exc)

Run it:

aioscraper scraper

What's happening?

1. @scraper registers the entry point; @scraper.pipeline registers a pipeline for RepoStats
2. schedule_request() queues a request and returns; the framework dispatches it when a slot frees up
3. Requests run concurrently up to the limit, so responses arrive in no particular order
4. parse_repo handles each response, on_failure handles each failure that was not retried
5. StatsPipeline.close() runs once at the end, after every request has finished

Retries are on by default. Rate limiting is not - turn it on, along with concurrency and timeouts, through environment variables before running this against a real API.

Examples

Runnable, commented scrapers live in examples/.

Why aioscraper?

vs Scrapy:

• Scrapy is built for HTML scraping with CSS/XPath selectors and website crawling
• aioscraper is optimized for API data collection (JSON, REST, microservices)
• Native asyncio (no Twisted), modern type hints, minimal footprint
• Easily embeds into existing async applications

vs httpx/aiohttp directly:

• Manual approach: you handle rate limiting, retries, queuing, concurrency, backpressure
• aioscraper: adaptive rate limits, priority queues, pipelines, middleware out of the box
• Declarative Request → callback → pipeline instead of imperative control flow

vs building custom async workers:

• Less boilerplate: focus on business logic, not infrastructure
• Production-ready components: EWMA+AIMD rate limiting, graceful shutdown, dependency injection
• Testable: explicit dependencies, no global state, easy mocking

When to use aioscraper:

• Collecting data from 100+ API endpoints
• Fan-out calls to microservices for data enrichment
• Queue consumers processing API scraping tasks
• API aggregation/monitoring pipelines
• High-throughput data collection jobs

Use Cases

1. E-commerce price monitoring

Poll 10,000 product API endpoints across multiple marketplaces:

• Adaptive rate limiting prevents bans
• Priority queue for trending products
• Pipeline aggregates prices → saves to DB → sends alerts on changes

2. Cryptocurrency data aggregation

Collect real-time prices from 20+ exchange APIs:

• Concurrent requests with per-exchange rate limits
• Built-in retry for transient failures
• Pipeline normalizes data formats → writes to time-series DB

3. Microservice data hydration

Your FastAPI app needs data from 50 internal services:

• Embed aioscraper in your async application
• Fan-out concurrent requests with backpressure control
• Middleware for auth, logging, circuit breaking

4. Queue-based scraping workers

Distributed architecture with Redis/RabbitMQ/SQS:

• Message queue publishes scraping tasks (URLs + params)
• aioscraper workers consume queue → fetch data → process
• Pipeline acknowledges messages after successful processing

5. Social media API aggregation

Aggregate user stats from Twitter, LinkedIn, GitHub APIs:

• Different rate limits per platform (adaptive throttling)
• Error callbacks for quota exceeded / auth failures
• Pipeline deduplicates → enriches → stores to database

6. Multi-source data snapshots

Collect point-in-time data from 500+ API sources simultaneously:

• Health monitoring: poll status endpoints of distributed services every minute
• Market data: snapshot prices from 200+ suppliers at exact intervals
• Analytics aggregation: fetch metrics from dozens of analytics APIs on schedule
• Concurrent execution with precise timing and automatic retries for failed sources

Performance

Benchmarks show stable throughput across CPython 3.11–3.14 (see benchmarks)

Documentation

Full documentation at aioscraper.readthedocs.io

Changelog

See CHANGELOG.md for version history and release notes.

Contributing

Please see the Contributing guide for workflow, tooling, and review expectations.

License

MIT License

Copyright (c) 2025 darkstussy