About
    • AMMDI is an open-notebook hypertext writing experiment, authored by Mike Travers aka mtraven. It's a work in progress and some parts are more polished than others. Comments welcome! More.
Search
Page Contents
Full
Page Tree

Blygger/ELI15

17 Oct 2026 - 17 Oct 2026
Open in Logseq

    • What is Blygger? (an explainer)

    • Written October 2026, when the protocol is at version 0.3 and there are about two dozen of these things in the world. Entirely LLM generated.

    • The one-sentence version

    • Blygger is a set of rules for publishing short writing on your own website in a way that other people's software can read, quote, and answer — without anybody needing a Facebook account, a server, or permission from a company.

    • Start with the problem

    • You already know two ways to put writing on the internet:
    • Social media. Easy, people are there, you get replies. But the company owns the floor you're standing on. They decide who sees your post, they can delete it, they can delete you, and if the company dies your writing dies with it. Everything you wrote on Vine is gone. Your Twitter threads from 2015 are probably broken.
    • A blog. You own it, it lasts, nobody can ban you. But it's lonely. A blog post is a 1,000-word production with a title and an opinion, and when you publish it, it just sits there. If someone writes a response on their blog, you find out by accident, if ever.
    • Blygger is an attempt at a third thing: writing that lives on your site like a blog, but that other people's software can follow, quote, and respond to like social media. Not a website you sign up for — a format, like RSS or email. Anybody can implement it. Nobody runs it.

    • What's actually on the wire

    • Here's the honest, unglamorous core: a blyg is a folder of files on a website, arranged in an agreed way. yoursite.com/blyg/ blyg.json ← "hi, I'm a blyg; here's my name and where my stuff is" feed.xml ← an ordinary RSS feed, so normal readers work items/index.json ← a list of everything you've ever published items/7c9wk2mh….json ← one post items/1vgtgz0g….json ← another post
    • That's it. No database, no login, no server code. If you can put files on the web — GitHub Pages, Neocities, a $3 host — you can run one. The whole protocol is agreements about what goes in those files.
    • Each post ("item") is JSON that looks roughly like: { "id": "7c9wk2mhq0v3xj8tn5rzfd41bg", "kind": "fragment", "version": 3, "content_md": "the post, as you typed it", "content_html": "the post, as a browser should show it", "changelog": [ {"version": 1, "at": "..."}, {"version": 2, "at": "..."} ] }

    • The five ideas that make it different

    • 1. Posts are small on purpose

    • The basic unit is a fragment — a thought, a paragraph, a quote with a comment. Not an essay. If you want something longer, you make a thread: a post built out of other posts, each of which is still its own real post with its own link. Think of it as building with Lego bricks instead of carving a statue.

    • 2. Nothing ever 404s

    • Every post gets a permanent 26-character id when it's created. The rule — and it's a hard rule, not a nice-to-have — is that a post's file returns something forever. Even if you delete the post. Deleting doesn't remove the file; it replaces it with a tombstone that says "this was withdrawn." Anyone who linked to you gets a straight answer instead of a dead link.
    • This sounds boring. It is the most radical thing in the protocol. The normal web rots; this one is designed not to.

    • 3. Editing is on the record

    • When you change a post, its version number goes up and the change is logged. The old text isn't published (that stays private by default — you're allowed to fix your typos in peace), but the fact that you edited is visible. No stealth edits.
    • If you want a specific version to be permanently readable — because someone's going to quote it and you want that quote to stay honest — you pin it. A pinned version gets its own permanent file. Pinning is one-way: once pinned, always available.

    • 4. Quoting copies the words, with a receipt

    • On the normal web, quoting is a link. The link says "go look over there," and when "over there" changes or vanishes, your quote is meaningless.
    • Blygger has transclusion: when you quote someone, your publishing software copies their actual words into your post, and attaches a note saying whose words, from which version, at which site. The copy doesn't change when they edit later — your quote is a photograph, not a window. And because the note names the exact version, any reader's software can check whether the original has moved on since.

    • 5. Replies live on your own site

    • There's no comment box. If you want to respond to someone, you publish a response on your own blyg — called a stub — which quotes them and is marked as "this is a reply to that." It appears in your feed, under your name. Then your software pings theirs to say "hey, you've been answered."
    • That ping is called a Webmention, and the clever part is how it's checked. When their site gets the ping, it doesn't just believe it — it goes and fetches your post and verifies your post really does name them. You can't fake a response, and you can't put words on somebody else's page. The spec says it plainly: "The stub is on the stubber's soapbox, under the stubber's identity, in the stubber's feed."
    • Nothing you do can make text appear on my site. Nothing I do can make text appear on yours. That single property removes most of what makes comment sections miserable.

    • Side by side with normal blogging

    • Normal blogBlygger
      Where it livesYour siteYour site
      Typical post1,000-word essay with a titleA paragraph; titles optional
      Who can read itHumans; RSS readers get a summaryHumans, RSS readers, and other blygs that can quote it precisely
      Dead linksDelete a post → 404 foreverDeleted posts leave a tombstone; links never break
      EditingSilently rewrite historyVersion number + changelog; stealth edits visible
      Quoting someoneA link that may rotA copy of their actual words, stamped with version and origin
      RepliesComment box (spam, moderation) or nothingReplies live on the replier's own site; you get a verified notification
      Can someone put text on your page?Yes, via commentsNever. Structurally impossible
      DiscoveryGoogle, luckA directory, plus "blogrolls" — lists of who you read
      What you needA blog hostAny web host that serves files
      Running cost$0–10/monthSame, often $0
      If the software diesExport and hopeThe files are the format; nothing to export
      Audience sizeEveryone on the webEveryone on the web — but the blyg-native part is ~2 dozen sites
      Algorithm deciding who sees youNone (that's the point)None

    • Advantages

    • You can't be deplatformed. It's your domain, your files. There's no account to suspend because there's no company.
    • It degrades gracefully. A blyg includes an ordinary RSS feed, so people with no idea what Blygger is can follow you in any feed reader. The fancy parts are additions, not requirements.
    • Links last. The "never 404" rule means a conversation from today still resolves in ten years, as long as your domain is alive.
    • Quotes are honest. Copy + version stamp + origin means a quote can be checked, and a quote of a deleted post doesn't become a mystery.
    • No comment section to moderate. Responses exist, but on other people's property. Nobody can dump abuse onto your page, so you never have to police one.
    • It's cheap and simple. Static files. No database to get hacked, no server to keep patched, no monthly bill beyond hosting.
    • Nothing decides what you see. Your reading list is exactly who you chose to follow, in the order they posted.
    • You can build on each other's work. You can take a pinned version of someone's post and fork it — continue or rework it, with the lineage recorded, like forking code.

    • Disadvantages

    • Be fair about these; they're real.
    • Almost nobody is there. As of October 2026 the public directory lists 23 blygs and 14 plain feeds. That's a village, not a town. If you want an audience, this isn't where it is yet.
    • Replies often go nowhere. Of those 23 blygs, only 12 can receive Webmentions. Reply to one of the other 11 and your response is published correctly and they simply never hear about it — because receiving notifications needs a running server, and a plain static site can't.
    • You need a website and some patience. There's no "sign up" button. In practice you need a domain, a host, and software that speaks the protocol — and the software is all young.
    • It's a moving target. The spec went from 0.1 to 0.3 in about a month, with 0.4 being planned. Things that are "not implemented yet" today may be normal next month, and anything you build has to keep up.
    • Quotes go stale by design. Because a quote is a copy, it can show an old version of someone's post. The protocol prefers this (an unchanging quote is honest about what you actually read) but it does mean a conversation can drift out of sync.
    • Your edit history is private unless you pin it. Good for your dignity, slightly awkward for anyone trying to cite you — they can only safely point at versions you deliberately froze.
    • Discovery is manual. No search, no recommendations, no trending. You find people through a directory, through blogrolls, or because somebody quoted them.
    • Clients are uneven. Every implementation supports a different slice of the spec, so what you publish and what a given reader shows you don't always match. (There's an automated conformance checker that regularly finds real sites failing real rules — including, at the time of writing, this one.)

    • Is it worth doing?

    • It depends what you want. If you want reach, no — 23 sites is not reach.
    • What it's actually good at is being a permanent, self-owned place to think in public that other people can nonetheless plug into. If you already write things down and you'd like them to still exist, still resolve, and occasionally get a real answer from a real person who went to the trouble of writing a response on their own site — that's the trade. Fewer people, but every one of them deliberate, and none of it rented.

    • Glossary

    • WordMeaning
      blygThe folder of files; also "a site that publishes one"
      itemOne post
      fragmentA plain, short item
      threadAn item built out of other items
      manifest (blyg.json)The file that says "I'm a blyg" and where everything is
      versionEdit counter on an item; every change bumps it
      pinFreezing one version so it's permanently readable and quotable
      transclusionQuoting by copying the words, stamped with version and origin
      stubA post that's marked as a reply to exactly one other post
      WebmentionThe ping that tells someone you replied to them
      withdrawal"Deleting" a post — leaves a tombstone, never a 404
      forkTaking a pinned version of someone's post and continuing it yourself
      originThe exact base URL that identifies a blyg (example.com/blyg/)
      blogrollA published list of the blygs you read