GitHub Trending Scraper — AI Agent & MCP Repos by Stars

Scrape trending GitHub repositories for any topic, with AI agents and MCP (Model Context Protocol) servers set up by default. Each row has the repo name and URL, stars, forks, open issues, topics, language, license, homepage and created / updated / pushed dates. It uses the official GitHub Search API, so there is no login, browser or proxy, and the default input finishes in under a minute.

Price: $1.00 per 1,000 repos + $0.001 per run. 50 repos cost $0.051.

Use cases

• Find breakout repos early: set createdAfter to the last 7 or 30 days and sort by stars to see the new projects everyone is starring.
• Track MCP servers and agent frameworks: a daily feed of new mcp-server, ai-agent and llm-agent repos for directories, newsletters and awesome-lists.
• Investors and analysts: monitor open-source momentum in agentic AI and build your own star history.
• Developer relations and competitive intel: watch repos in your category, with language and license filters.
• Any other niche: change the topics or keywords, e.g. rust, rag, computer-vision.

What does it do?

It runs one GitHub search per topic (default: mcp-server, mcp, model-context-protocol, ai-agent, ai-agents, llm-agent, agentic-ai, autonomous-agents) plus any keywords, merges and de-duplicates the results, records which search matched each repo (matched_queries) and sorts by stars, last update or creation date. Adding createdAfter gives you new repos ranked by stars, which is the closest thing to "GitHub trending" for a topic. A GitHub token is optional and only raises the rate limit.

Input

Field	What it does
topics	GitHub topics to search (results are merged and de-duplicated)
keywords	Extra words to match in repo names and descriptions
createdAfter	Only repos created on or after this date (YYYY-MM-DD)
minStars	Minimum number of stars
language	Only repos in this language, e.g. Python
sort	stars, updated or created
maxItems	Maximum number of repos to return (default 50)
githubToken	Optional GitHub token for a higher rate limit (stored as a secret)

Example input (the run the output below comes from):

{
  "topics": ["mcp-server", "ai-agent"],
  "createdAfter": "2026-10-01",
  "minStars": 20,
  "maxItems": 50
}

Example output

Real items from a run on 2 November 2026 (new mcp-server and ai-agent repos created since 1 October, ranked by stars):

[
  {
    "id": 1414685342,
    "full_name": "onehumanbeing/esp32-iphone-agent",
    "html_url": "https://github.com/onehumanbeing/esp32-iphone-agent",
    "description": "Let an AI model use a real iPhone with a $7 ESP32: AirPlay eyes, USB HID hands, any OpenAI-compatible vision model. No Mac, no developer account, no jailbreak.",
    "stargazers_count": 4219,
    "forks_count": 565,
    "open_issues_count": 16,
    "language": "C",
    "topics": [
      "ai-agent",
      "airplay",
      "automation",
      "computer-use",
      "esp32",
      "esp32-s3",
      "iphone",
      "llm",
      "usb-hid"
    ],
    "created_at": "2026-10-11T19:10:25Z",
    "updated_at": "2026-11-02T09:26:01Z",
    "pushed_at": "2026-10-28T07:40:06Z",
    "license": "MIT",
    "owner_login": "onehumanbeing",
    "owner_type": "User",
    "homepage": "https://handsy.tech?ref=github",
    "archived": false,
    "fork": false,
    "matched_queries": [
      "topic:ai-agent"
    ],
    "_meta": {
      "actor": "github-ai-agent-tracker",
      "schemaVersion": "1.0",
      "scrapedAt": "2026-11-02T09:44:34.307Z"
    }
  },
  {
    "id": 1418870464,
    "full_name": "Graphicle-AI/graphicle",
    "html_url": "https://github.com/Graphicle-AI/graphicle",
    "description": "Code graph and memory for your coding agents, in one Rust program. For Claude Code, Codex, Cursor and any MCP client.",
    "stargazers_count": 492,
    "forks_count": 35,
    "open_issues_count": 4,
    "language": "Rust",
    "topics": [
      "agent-memory",
      "ai-agents",
      "claude-code",
      "code-analysis",
      "code-graph",
      "code-intelligence",
      "code-search",
      "codex",
      "coding-agents",
      "context-engineering",
      "cursor",
      "developer-tools",
      "graphrag",
      "knowledge-graph",
      "local-first",
      "mcp",
      "mcp-server",
      "model-context-protocol",
      "rust",
      "tree-sitter"
    ],
    "created_at": "2026-10-14T12:25:36Z",
    "updated_at": "2026-11-02T08:26:47Z",
    "pushed_at": "2026-11-02T08:20:42Z",
    "license": "Apache-2.0",
    "owner_login": "Graphicle-AI",
    "owner_type": "Organization",
    "homepage": "https://www.karlex.ai/graphicle",
    "archived": false,
    "fork": false,
    "matched_queries": [
      "topic:mcp-server"
    ],
    "_meta": {
      "actor": "github-ai-agent-tracker",
      "schemaVersion": "1.0",
      "scrapedAt": "2026-11-02T09:44:40.305Z"
    }
  }
]

Export as JSON, CSV, Excel or HTML, or pull the data through the Apify API, webhooks, Make, Zapier, n8n or the Apify MCP server.

How much does it cost?

Pay per event, no subscription:

• $0.001 per run (actor start)
• $0.001 per repo returned ($1.00 per 1,000)

Examples: 50 repos cost $0.051; a daily run of 50 repos costs about $1.53 a month. You pay nothing for results on a failed run: if no data is found, the run fails with a clear message instead of returning an empty "success".

See the trends without running anything

Agent Data Co publishes a free daily momentum leaderboard of the fastest-rising AI repos and MCP servers, plus daily AI ecosystem trends. Use this Actor when you need the raw data for your own topics.

FAQ

Is this allowed? It uses the official GitHub REST API under GitHub's API terms and returns public repository metadata only. No personal data is collected.

How do I build a star history? Schedule a daily run (Apify Console → Schedules) and keep each dataset, or send it to Google Sheets or a database. Comparing stargazers_count across days gives star growth per repo.

Something broke? Open an issue on the Issues tab. The Actor is tested every day.

---

More AI-ecosystem Actors from Agent Data Co

Built and maintained by Agent Data Co. Our Actors collect public AI-ecosystem data and work well together on a daily schedule:

• Show HN Scraper (Hacker News AI launches): new AI launches on Hacker News with points and comments
• MCP Servers Scraper (official MCP registry): every server in the official MCP registry
• Hugging Face Scraper: trending models, datasets and Spaces
• LLM Pricing Scraper (OpenRouter): 400+ LLM prices with price-change detection
• SaaS Pricing Page Monitor: plan and price changes on any pricing page
• App Store Reviews Scraper: iOS reviews for any app, by country