Unified search engine for userscripts — query Greasy Fork, Sleazy Fork, GitHub, and OpenUserJS from a single interface.

Live: https://sysadmindoc.github.io/UserScriptHunt/

ScriptHunt is a zero-dependency, single-file HTML webapp that searches every major userscript repository in parallel and merges the results into one unified, deduplicated feed. No backend required — runs entirely in the browser, deployable as a static page on GitHub Pages.

Local:

git clone https://github.com/SysAdminDoc/UserScriptHunt.git
cd UserScriptHunt
# Open directly — no build step, no dependencies
open userscript-search.html

GitHub Pages:

1. Push to a GitHub repository
2. Go to Settings → Pages → Source → Deploy from branch (main, root)
3. Access at https://yourusername.github.io/UserScriptHunt/

Direct link with pre-filled search:

userscript-search.html?q=youtube+enhancer

Greasy Fork / Sleazy Fork

• Endpoint: https://api.greasyfork.org/en/scripts.json?q={query}&page={n}
• Direct fetch with native CORS (Access-Control-Allow-Origin: *)
• Largest userscript repository (~100,000+ scripts)
• 100 results per page
• Full metadata: daily/total installs, good/bad ratings, fan score, version, license, code URL

GitHub

• Endpoint: https://api.github.com/search/repositories?q={query}+userscript
• Searches repos tagged with userscript/tampermonkey/greasemonkey keywords
• Unauthenticated rate limit: 10 requests/minute for search
• Returns stars, forks, language, license, owner info

OpenUserJS

• Endpoint: https://openuserjs.org/?q={query}&orderBy=installs&orderDir=desc&p={n}
• HTML response parsed via DOMParser (no JSON API available)
• Routed through CORS proxy (allorigins.win with codetabs fallback)
• ~12,000 scripts indexed

┌─────────────────────────────────────────────────────────────────┐
│                        User Search Input                        │
│                     (450ms debounce / Enter)                    │
└──────────────────────────┬──────────────────────────────────────┘
                           │
              ┌────────────┼────────────┬────────────┐
              ▼            ▼            ▼            ▼
     ┌──────────────┐ ┌──────────┐ ┌────────┐ ┌───────────┐
     │ Greasy Fork  │ │ Sleazy   │ │ GitHub │ │ OpenUserJS│
     │  (Direct)    │ │ (Direct) │ │(Direct)│ │  (Proxy)  │
     └──────┬───────┘ └────┬─────┘ └───┬────┘ └─────┬─────┘
            │              │           │             │
            └──────────────┴─────┬─────┴─────────────┘
                                 │
                    Promise.allSettled()
                                 │
                                 ▼
                  ┌──────────────────────────┐
                  │  Normalize + Deduplicate  │
                  │   (ID + name::author)     │
                  └─────────────┬────────────┘
                                │
                                ▼
                  ┌──────────────────────────┐
                  │    Sort + Render Cards    │
                  │  (Grid with source badges)│
                  └──────────────────────────┘

All source requests fire in parallel. Results render progressively as each source responds — fast sources appear immediately while slower ones are still loading. Failed sources display an error chip but never block other results.

Click the source chips below the search bar to enable/disable sources at runtime. Active sources have a bright dot and highlighted border. Disabled sources are skipped entirely (no requests sent).

Default state:

• Greasy Fork: Enabled
• Sleazy Fork: Disabled (opt-in for adult content)
• GitHub: Enabled
• OpenUserJS: Enabled

The app uses CORS proxy services with automatic sequential fallback, but only for OpenUserJS (HTML scraping). Greasy Fork, Sleazy Fork, and GitHub all have native CORS support and are fetched directly with no proxy.

Proxy chain for OpenUserJS:

1. api.allorigins.win/get (primary — wraps HTML in JSON)
2. api.codetabs.com (fallback — passes through raw content)

To use a custom proxy, modify the fetchViaProxy function in the script:

// Example: Cloudflare Worker proxy
const resp = await fetch(`https://your-proxy.workers.dev/?url=${encodeURIComponent(url)}`);

GitHub results seem broad / not all are userscripts GitHub source searches repositories (not individual files) using keyword matching against userscript OR tampermonkey OR greasemonkey. This casts a wider net but may include related tools and libraries alongside actual userscripts. Unauthenticated code search (which would filter to .user.js files specifically) is restricted by GitHub's API.

OpenUserJS returns no results or fails OpenUserJS results depend on CORS proxy availability. If both proxies are down or rate-limited, OpenUserJS will show a "Failed" status chip. Results from other sources will still display normally.

Rate limits GitHub enforces 10 search requests/minute for unauthenticated users. Heavy pagination or rapid searching may trigger a 403 response — the app will surface this as an error chip. Greasy Fork has no documented rate limits but self-throttles via 100-item pagination.

Can I self-host a CORS proxy? Yes. Deploy a Cloudflare Worker (free tier: 100K requests/day) as a simple CORS proxy and update the fetchViaProxy function. This eliminates reliance on third-party proxy services.

• Single-file HTML — no build step, no bundler, no framework
• Vanilla JavaScript (ES2022) — async/await, Promise.allSettled, AbortController, DOMParser
• CSS Custom Properties — full theming via variables
• Google Fonts — JetBrains Mono (logo/monospace) + Outfit (UI)
• CORS Proxy — allorigins.win/codetabs fallback chain, only used for OpenUserJS HTML scraping
• No localStorage — stateless, no tracking, no cookies

• Userscript.Zone integration (HTML scraping — would add 10+ aggregated sub-sources)
• ScriptCat store integration
• GitHub authenticated code search (.user.js file-level search)
• Metadata block preview (parse @match, @grant, @require from script headers)
• One-click install URLs (direct .user.js links for Tampermonkey/Violentmonkey)
• Search history / saved searches (sessionStorage)
• Filter by target site (e.g. show only scripts for youtube.com)
• Dark/OLED theme toggle

Issues and PRs welcome. If adding a new source, follow the existing adapter pattern:

1. Create a searchSourceName(query, page) async function
2. Return an array of normalized result objects matching the schema
3. Add the source to the SOURCES config object
4. Wire it into executeSearch()

MIT License — see LICENSE for details.