Anaconda / Conda Package Scraper

Scrape conda package data from anaconda.org — versions, total downloads, license, supported platforms, package types, maintainer channel and project links — with no API key, no login and no proxy required.

Point it at channel/name pairs like conda-forge/numpy or bioconda/samtools, or search the whole registry by keyword and get one clean, flat row per package.

---

What you get

One dataset item per conda package:

Field	Description
name	Package name, e.g. numpy
full_name	channel/name, e.g. conda-forge/numpy
channel	Owning channel / organisation (conda-forge, bioconda, anaconda, …)
summary	Short description (capped at 1,000 chars)
description	Long description (capped at 3,000 chars)
license	SPDX-style license string, e.g. BSD-3-Clause
latest_version	Newest published version
version_count	Number of published versions
total_downloads	Sum of downloads across every uploaded artifact
file_count	Number of uploaded build artifacts
platforms	Unique subdirs, e.g. linux-64, osx-arm64, win-64
package_types	e.g. conda, pypi, standard
home_url, doc_url, dev_url, html_url	Project homepage, docs, repo, anaconda.org page
created_at, modified_at	ISO-8601 UTC timestamps
watchers	Watcher count on anaconda.org
source, scraped_at	Provenance metadata

Every field is nullable — community channels publish very sparse metadata and the scraper never invents values.

Sample output

{
  "name": "numpy",
  "full_name": "conda-forge/numpy",
  "channel": "conda-forge",
  "summary": "The fundamental package for scientific computing with Python.",
  "license": "BSD-3-Clause",
  "latest_version": "2.5.1",
  "version_count": 115,
  "total_downloads": 145216988,
  "file_count": 3739,
  "platforms": ["linux-64", "linux-aarch64", "linux-ppc64le", "osx-64", "osx-arm64", "win-64"],
  "package_types": ["conda"],
  "home_url": "http://numpy.org/",
  "doc_url": "https://numpy.org/doc/stable/",
  "dev_url": "https://github.com/numpy/numpy",
  "html_url": "http://anaconda.org/conda-forge/numpy",
  "created_at": "2020-06-06T23:22:39.637000+00:00",
  "modified_at": "2026-07-05T00:25:51.477000+00:00",
  "watchers": 1,
  "source": "anaconda",
  "scraped_at": "2026-07-28T12:13:48.770763+00:00"
}

---

Input

Field	Type	Default	Description
mode	select	packages	packages = fetch details for given channel/name pairs. search = keyword search the registry.
packages	array	["conda-forge/numpy", "conda-forge/pandas"]	Packages as channel/name (full anaconda.org URLs also work).
queries	array	–	Search keywords, e.g. ["pytorch"].
maxItems	integer	100	Hard cap on dataset items (max 1,000).

Example: track your dependency stack

{
  "mode": "packages",
  "packages": [
    "conda-forge/numpy",
    "conda-forge/pandas",
    "conda-forge/scikit-learn",
    "bioconda/samtools",
    "pytorch/pytorch"
  ],
  "maxItems": 100
}

Example: discover packages by keyword

{
  "mode": "search",
  "queries": ["pytorch", "bioinformatics", "geospatial"],
  "maxItems": 300
}

---

Use cases

• Data-science tooling — keep an internal catalogue of the conda packages your notebooks and pipelines depend on, with live version and platform coverage.
• Dependency audits — flag packages whose latest version is stale, whose license changed, or that dropped a platform such as osx-arm64 or win-64.
• Package discovery — search the registry for a topic and rank candidates by download volume before committing to a dependency.
• Open-source intelligence — benchmark adoption of libraries across conda-forge, bioconda and vendor channels using download counts and release cadence.

---

Pricing

Roughly $0.50 per 1,000 packages. Runs on 512 MB and hits a public JSON API directly, so there is no browser overhead, no proxy cost and no rate-limit dance.

Speed & reliability

• Fully async, batched up to 50 packages in flight.
• Chrome 124 TLS fingerprint via curl_cffi, with a stdlib urllib fallback.
• Retries with exponential backoff on 429/5xx; missing packages are logged and skipped rather than crashing the run.
• Datacenter-clean: works on Apify's default datacenter IPs, no residential proxy needed.

Notes & limits

• Data comes from the public api.anaconda.org endpoints; private channels are not accessible.
• Search returns registry index records, which omit per-file data — in search mode file_count is null and total_downloads falls back to the channel-reported aggregate.
• Only public metadata is collected. No personal data, no login, no scraping of gated content.

Related actors

• PyPI Scraper
• npms Search Scraper
• crates.io Scraper
• Hex.pm Scraper

---

Keywords: anaconda scraper, conda package scraper, anaconda.org API, conda-forge data, conda package downloads, package registry scraper, dependency audit, Python package metadata.