SpinSignal
About

An independent OSINT research umbrella.

SpinSignal uses software to build archives of subjects that are changing faster than the record of them can survive. Each subject gets its own project, called a vertical, with its own way of collecting and its own website.

Rating plate 1 operator · 0 funders · 4 verticals
Run by
One person With a lot of help from software.
Funding
None Nothing is commissioned or sponsored.
AI models
Local They read and tag the documents.
Published as
Snapshots Fixed web pages, not live feeds.

What this is

3 points

Each vertical is an automated collector pointed at one fast-changing topic. Saving things comes first. Making sense of them is a distant second.

  1. It collects

    It gathers material from a fixed list of sources, on a schedule. It keeps the full text, makes it easy to browse, and builds a website from the result.

  2. It saves things early

    The most valuable thing it does is grab material before it's edited, put behind a paywall, moved or taken down. Where possible, it also sends a copy to the Internet Archive's Wayback Machine at the moment it's collected.

  3. Same method, different subjects

    The verticals share a method, not a subject. What they have in common is an original record that's disputed, easy to lose, and being rewritten while people are still reading it.

What this is not

4 things

Four things this isn't, said plainly.

  1. Not a news outlet.

    Nothing here is reported by journalists. There's no newsroom, no editor and no house view. Where a site shows a summary, software wrote it, and you can check it against the archive it came from.

  2. Not affiliated.

    SpinSignal isn't connected to any organisation, government, institution, publisher or funder. There's no funding to disclose, because there isn't any. Nothing here is commissioned or sponsored.

  3. Not an intelligence product.

    This is research by one person and some software. It isn't verified reporting, and it's not a sound basis for decisions about operations, publishing, health or money. Use the archive to find original sources, then read those.

  4. Not neutral by design.

    Choosing which sources to collect is an editorial choice, and every vertical makes it. Each one publishes its source list, so you can judge the choices yourself.

Live or archived

2 kinds

Every vertical is one of two kinds, and it says which on its own front page.

Live tracker

Still collecting. It checks its sources on a schedule, and rebuilds its website from whatever it has stored so far.

Point-in-time archive

Stopped. Collecting has ended, and the website shows what was held on one particular date.

Worth being clear about

Both kinds are static websites, which means even a live tracker's site is a snapshot. It shows what was stored the last time the site was rebuilt, not what's stored right now. Each vertical carries a provenance stamp, a label showing when its site was built. A page that looks live and a page that's up to date are different things, and that gap is one of the ways this can go wrong. It has its own page.

Who runs it

2 points

One person, working independently, with a lot of help from software.

  1. Local AI models

    The AI models that read and tag the documents run locally, not on someone else's service. That gives full control, privacy, and the freedom to explore without restrictions or limitations.

  2. A limit that comes with it

    The trade-off is that local models are limited in what they do well, which Failure Modes explains.