Continue.dev Hub Catalog Scraper

Scrapes the complete Continue.dev hub catalog — the canonical marketplace for AI coding assistant configurations. Extracts rules, prompts, models, documentation packages, and MCP server blocks with full config YAML, star counts, and author details.

What it scrapes

The Continue.dev hub (https://continue.dev/hub) is an open marketplace where teams publish their AI coding assistant configurations. This actor collects all public packages across these categories:

Type	Examples
rules	Coding standards, project conventions (e.g. Angular Rules, Next.js Rules)
prompts	Reusable prompts for common workflows
models	Model configuration blocks (OpenAI GPT-4o, Claude, Gemini, Llama, etc.)
docs	Framework documentation packages
blocks	Composite configuration blocks
mcp-servers	MCP server integration configs

Use cases

• Enterprise config auditing — inventory all publicly-shared Continue.dev configs to benchmark your team's setup against community standards
• AI dev-tool intelligence — track adoption of models, coding standards, and MCP integrations across the developer community
• Config discovery — find the most-starred rules/prompts for a specific framework before building your own
• Research — analyze trends in AI coding assistant configuration patterns

Output fields

Field	Type	Description
item_id	string	Unique ID in author/slug format (e.g. starter/angular-rules)
item_type	string	Package type: rules, prompts, docs, models, blocks, mcp-servers
author	string	Owner/author slug
name	string	Display name of the package
tagline	string	Short description
is_official	boolean	Whether this is a verified/official package
star_count	integer	Number of stars
use_count	integer	Number of installs/uses
raw_markdown	string	Full YAML/Markdown content of the package config
profile_url	string	Direct URL to the package page
scraped_at	string	ISO-8601 scrape timestamp

Input parameters

Parameter	Type	Default	Description
maxItems	integer	50	Maximum number of items to scrape
itemTypes	array	all types	Filter to specific types: rules, prompts, docs, models, blocks, mcp-servers

Sample output

{
  "item_id": "starter/angular-rules",
  "item_type": "rules",
  "author": "starter",
  "name": "Continue * starter/angular-rules",
  "tagline": null,
  "is_official": true,
  "star_count": 9,
  "use_count": 0,
  "raw_markdown": "---\nname: Angular Rules\n---\n\n- You are an Angular developer\n- Use Angular CLI for project scaffolding\n...",
  "profile_url": "https://www.continue.dev/starter/angular-rules",
  "scraped_at": "2026-06-12T14:30:00.000Z"
}

Technical notes

• No authentication required — all hub packages are publicly accessible
• No proxy required — datacenter IPs work fine
• Typical full run: ~100-120 items across all types, completes in under 2 minutes
• Data is extracted from Next.js RSC streaming payloads embedded in the HTML