Signal Collective
Hong Kong Book a call
Signal Collective The Library Research
We find where your growth is · Then we build it

Research · What changed, and what people are saying, before you decide. · a tool

Search the web without being tracked

The open-source project searxng, by searxng

“SearXNG is a free internet metasearch engine which aggregates results from various search services and databases. Users are neither tracked nor profiled.” — the project's own words, on GitHub

What it does
It's a search engine you run yourself, that pulls results from several other search engines at once and returns one combined list — without logging what you searched for or building a profile of you. Point your AI at it instead of a single search API and it gets the breadth of results a person would get by checking several engines by hand, in one call. Because you're the one hosting it, no outside company sees your queries, and you decide which sources it checks and how much weight each one gets. It's the kind of thing you set running once and mostly forget.
Replaces
a paid search API
For
Anyone who wants their AI's web searches to go through something they control, not a third-party search API logging every query.
Not for
Not for someone without a machine they're willing to leave a small server running on — this is self-hosted only, there is no hosted version to sign up for.
Setup
15 min · Claude Code, Kimi, Gemini CLI or Codex · docker, a machine you can leave running (a small VPS or a home server)
Twenty tabs, open One filter One page, with sources public pages · robots respected WHAT CHANGED two you open yourself

Get it running

Paste this into the AI that runs on your computer. It does the install, checks it works, and tells you what to type first. If anything fails, paste the error back to it.

Paste into Claude Code
Install searxng (https://github.com/searxng/searxng) on this computer and get it working for me. I am not a developer; explain each step in one line as you go and never skip one. You are Claude Code, running on my machine.

1. Check what this machine already has (git, Python or Node as the project needs, Docker if the README says so). Tell me anything missing and install it, asking me before anything that needs my password.
2. Clone https://github.com/searxng/searxng into ~/tools/searxng and follow the README's install exactly.
3. Configure it. If it needs an API key or a login, stop and ask me for it; never guess one and never store it anywhere except where the README says.
   Tool-specific notes: Run it with `docker run --name searxng -d -p 8888:8080 -v "./config/:/etc/searxng/" -v "./data/:/var/cache/searxng/" docker.io/searxng/searxng:latest`, and create a `settings.yml` in `./config/` first.
Smoke-test with `docker container list` to confirm it's up, then visit `http://localhost:8888` in a browser and run one search.
If Docker Hub is rate-limiting pulls, use the mirror image `ghcr.io/searxng/searxng` instead.
Don't expose port 8888 to the open internet without your own authentication in front of it — there's no login built in.
4. Run the smallest test the README gives, and show me the output.
5. When it works, tell me what to type first, in one line, for this job: Run one real search.
If anything fails, show me the exact error and fix it before going on.

What to point it at first

1

Run one real search

Search for something you'd normally google and check the results actually came back from more than one underlying engine, not just one.

2

Point your AI at it

Configure your AI tool to use this instead of its default search and confirm it's genuinely hitting your instance by checking your own container's logs for the request.

3

Turn a source off

Open its settings, disable a source you don't trust or don't need, and confirm your next search no longer includes it.

What it must never do unattended

  • Never publish what it found without opening the sources yourself.
  • Never point it past a login or a CAPTCHA. Public pages only, and it respects robots.txt.
  • Never leave it reachable on the open internet without your own authentication in front of it — there is no login built in.

Who made it

searxng/searxng on GitHub, under the AGPL-3.0 licence. 36,689 stars, checked 9 September 2026. Last change 8 September 2026. We did not write it; we checked it, and wrote this page so you can use it.

Also in Research

Twenty tabs, sometime One page, every Monday MONDAY BRIEF competitor onecompetitor twocompetitor three every claim · a source · two you check
002

Weekly Research Brief

Once a week, what changed with your competitors and in your niche. One page, every line sourced.

15 min · Claude, ChatGPT or Gemini · three competitor namesOpen
Twenty tabs, open One filter One page, with sources public pages · robots respected WHAT CHANGED two you open yourself
Tool · Research · crawl4ai

Turn any web page into notes your AI can use

Give it a web address, or a list of them, and it hands back clean, plain-text versions of those pages — headings and paragraphs stripped of surrounding code — so your AI can read them.

20 min · your AI on your computer · a Python 3.10+ environment, playwright's browser binaries (installed via crawl4ai-setup), docker (optional, for server mode)Open
Twenty tabs, open One filter One page, with sources public pages · robots respected WHAT CHANGED two you open yourself
Tool · Research · firecrawl

Pull what you need off any website

Give it a web address and it hands back the page as clean data — plain-text markdown, or JSON shaped the way you ask for — so your AI can read it without wading through a page's raw code.

15 min · your AI on your computer · an API key from firecrawl.dev (or a self-hosted instance), python or node to call itOpen
Twenty tabs, open One filter One page, with sources public pages · robots respected WHAT CHANGED two you open yourself
Tool · Research · gpt-researcher

A research report in minutes, with sources

You give it a question and it goes and researches it properly — pulling from many web sources at once, reading past what any single article says, and writing up a full report with citations you can check yourself.

20 min · your AI on your computer · python 3.11+, an OpenAI API key, a Tavily API key (for its web search), docker (optional, for the bundled web app)Open

What's in the library is how we work.

If you're launching or growing a brand across Asia and the West, a call is where we work out whether it's a fit.