Data
Plugin
insane-search
Impossible is nothing. If it's public, insane-search gets in. A resilient public-page reader for Claude Code. No API keys, no proxy setup.
Plugin Signal69/100Solid
maintained98
capable14
documented100
adopted71
Install
$ npx -y skills add fivetaku/insane-search --agent claude-code
Run the curl in your terminal, the rest in Claude Code.
Stats
2,264
Stars
274
Forks
Active
Maintenance
Python
Language
MIT
License
3d ago
Last commit
4mo ago
Created
Repo: fivetaku/insane-search
What's inside
README
1English | νκ΅μ΄ | δΈζ | ζ₯ζ¬θͺ | EspaΓ±ol
insane-search
Impossible is nothing. If it's public, insane-search gets in.
A resilient public-page reader for Claude Code. No API keys, no proxy setup.
β‘ Install
/plugin marketplace add https://github.com/fivetaku/gptaku_plugins.git
/plugin install insane-search@gptaku-plugins
/reload-plugins
No commands to learn. Ask Claude Code normally β insane-search kicks in when a fetch gets blocked.
π¬ Try it
Just ask normally β insane-search kicks in when a fetch gets blocked:
"Find what people are saying about Claude Code on Reddit and summarize the top threads." "Search X for posts about insane-search." "Summarize this YouTube video."
Expected: Claude reaches each site's public route β Reddit's feed, X via oEmbed, YouTube captions β with no login and no API key, and returns usable text, where the same request returns "I can't access that" without the plugin.
π Works on
X Β· Reddit Β· YouTube Β· Hacker News Β· Naver Β· Coupang Β· LinkedIn Β· Medium Β· Substack Β· arXiv Β· GitHub Β· Stack Overflow Β· Bluesky Β· Mastodon β plus any site with a public page, feed, or /rss. Full platform list & methods β PLATFORMS.md.
βοΈ Why it gets through
- It escalates, never pre-judges β public API readers β syndication gateways β TLS impersonation β a real headless browser, trying each route until one works.
- It looks human β builds a full browser identity (real TLS fingerprint, cookie warming, referer chain), not just a swapped User-Agent.
- Zero setup β auto-installs what it needs (
curl_cffi,yt-dlp, β¦) on first use. No API keys, no signup.
π Default Claude Code vs + insane-search
| When you hit⦠| Claude Code alone | + insane-search |
|---|---|---|
A 403 / WAF-blocked page | β gives up | β escalates until one route gets through |
| Platform content (X, Reddit, HN) | β often blocked or empty | β public API readers + syndication |
| An anti-bot / CAPTCHA challenge | β stops | β TLS impersonation, then a real headless browser |
| A media page (YouTube, 1,800+ sites) | β no transcript | β yt-dlp metadata + captions |
Missing tools (curl_cffi, yt-dlp) | β | β auto-installs on first use |
| API keys / signup | β | β none |
| A login wall or paywall | β | β stops here, and says so (see Boundaries) |
The differentiating row is the one thing the default fetcher can't do: it keeps trying public routes until one works.
β¨ What just happened
| Without it | With insane-search |
|---|---|
Ordinary fetch hits 403 and Claude says it can't read the page | The plugin picks a public-access route and returns usable text |
| You manually try mirrors, archives, mobile URLs | Fallbacks escalate automatically, public-only |
| A blocked page is a dead end | Metadata and structured data still yield titles, prices, summaries |
π How it works
Each phase runs only if the previous one fails or detects a blocking signal:
- Phase 0 β special public endpoints (platform APIs, feeds) it can't discover generically
- Phase 1 β lightweight probes: public API readers, syndication gateways, mobile /
.json//rssURL variants - Phase 2 β TLS impersonation (curl_cffi: safari β chrome β firefox) with a full browser identity
- Phase 3 β a real headless browser for JS-rendered pages
- Exit β a login or paywall is detected: it reports "authentication required" rather than pretending
Every response is also scanned for OGP / JSON-LD, so even partial pages yield titles, summaries, and prices.
π Boundaries
insane-search is a reader for public content, not a way around authentication.
- Reaches what's available through public pages, public APIs, feeds, archives, and a browser's public responses.
- Stops at logins and paywalls β it reports
authentication requiredinstead of trying to defeat them. - Never logs in as you and never stores or transmits credentials.
- All routes use no-auth public endpoints and standard, documented techniques.
- You are responsible for lawful, authorized use β each site's Terms of Service,
robots.txt, rate limits, and applicable law. This is a dual-use developer tool; see DISCLAIMER.md.
License
MIT β see DISCLAIMER.md for acceptable-use and user-responsibility terms.
Part of gptaku-plugins β Claude Code plugins that break through the walls everything else stops at.
File tree
65fivetaku/insane-searchView on GitHub
.claude-plugin/
plugin.json
.gitignore
assets/
hero.png
pipeline.png
CHANGELOG.md
DISCLAIMER.md
LICENSE
PLATFORMS.md
README.es.md
README.ja.md
README.ko.md
README.md
README.zh.md
setup/
gptaku-update-check.cjs
setup.sh
skills/
insane-search/
engine/
__init__.py
__main__.py
bias_check.py
content_safety.py
executor.py
fetch_chain.py
learning.py
observations_log.py
phase0.py
safety.py
templates/
.gitignore
nodriver_fetch.py
package.json
patchright_fetch.py
playwright_mobile_chrome.js
playwright_real_chrome.js
tests/
test_smoke.py
test_t1_retry.py
test_t2_rescue.py
test_t3_markdown.py
test_t4_maincontent.py
test_t5_pdfplumber.py
test_t6_differential.py
test_t7_browser_gate.py
test_u1.py
test_u4.py
test_u5.py
test_u7.py
test_u8.py
test_u9.py
transport.py
url_transforms.py
validators.py
waf_detector.py
waf_profiles.yaml
references/
cache-archive.md
fallback.md
jina.md
json-api.md
media.md
metadata.md
naver.md
playwright.md
public-api.md
rss.md
tls-impersonate.md
twitter.md
SKILL.md
tests/
coverage_battery.py
live_benchmark/
bench.pyFAQ
insane-search is a Claude Code plugin with 1 hand-picked skill for data work, indexed on Flowy. Install it with the command on its page. It includes insane-search. Its skills do not fire on their own yet. Request auto-invocation to have Flowy route them as you prompt. Free and open source.
- What is insane-search?
- Impossible is nothing. If it's public, insane-search gets in. A resilient public-page reader for Claude Code. No API keys, no proxy setup.
- How do I install insane-search?
- Run these in Claude Code: npx -y skills add fivetaku/insane-search --agent claude-code. Then prompt normally.
- Does insane-search auto-invoke its skills?
- Not yet. It is indexed on Flowy as a plain plugin. Request auto-invocation on its page and Flowy will route its skills for you as you prompt.
- Is insane-search free?
- Yes. Flowy is free and open source, with nothing gated. You can read every skill in full before you install.