GitHub and HuggingFace AI Research Monitor

Track trending AI repositories, models, datasets, and papers from GitHub and HuggingFace in a single Apify actor run. Built for researchers, ML engineers, and anyone who needs a daily pulse on the AI ecosystem without manually checking multiple sites.

What it does

This actor pulls data from two major platforms:

GitHub

• Scrapes the GitHub Trending page (daily, weekly, or monthly) to find repos gaining momentum
• Uses the GitHub Search API to find repos tagged with AI-related topics (machine-learning, deep-learning, llm, nlp, etc.)
• Deduplicates across both methods so each repo appears once in the output

HuggingFace

• Trending models via the HuggingFace API (sorted by trending)
• Trending datasets
• Trending daily papers with titles, authors, and abstracts

Input

Field	Type	Default	Description
sources	array	["github", "huggingface"]	Which platforms to monitor
max_items_per_source	integer	50	Max items per source category (1-500)
time_period	string	"daily"	Trending window for GitHub (daily/weekly/monthly)
github_topics	array	[]	Custom topics to filter GitHub repos (defaults to AI topics)
huggingface_categories	array	[]	Custom categories/tags for HuggingFace filtering
include_details	boolean	true	Fetch additional details per item

Output

Each item in the dataset contains:

• type - Item type: github_repo, hf_model, hf_dataset, or hf_paper
• source - Platform: github or huggingface
• url - Direct link to the item
• crawled_at - ISO timestamp of when the data was fetched
• Plus type-specific fields (stars, downloads, likes, tags, authors, etc.)

Usage

Run locally with the Apify CLI:

cd github-huggingface-research-monitor
python -m venv .venv
PYTHONPATH="" .venv/Scripts/pip install -r requirements.txt
PYTHONPATH="" apify run --purge

Deploy to Apify platform:

$PYTHONPATH="" apify push

Rate limits

• GitHub Search API: 60 requests/hour without authentication. The actor uses one search call plus one trending page scrape, well within limits.
• HuggingFace API: No authentication needed, uses reasonable request rates.

Use cases

• Daily AI research briefings
• Competitive monitoring for ML teams
• Feeding into newsletters or Slack bots
• Building trend dashboards
• Identifying rising projects early

Changelog

0.1 - Initial release. GitHub trending + search, HuggingFace models/datasets/papers.