Vol. I · Mon, 16 Nov 2026Open 24 hours · Feeding: continuous · Please do not tap the glassPatrons☕ Keepers' coffee
Staff only beyond this point

The Trap Room

Please lower your voice. Some of them are still in the net.

Follow a bot through the maze, live →
How the trap works
  1. The site's robots.txt says, plainly: Disallow: /trap/
  2. Every page hides a link to /trap/ that no human can see or click.
  3. Anything that walks through it has ignored the sign and come in anyway.
1Caught today
415Caught in 30 days
72Species ever caught
0Escapes prevented

The Descent

Follow one →LIVE

A cutaway of the Labyrinth, right now. Every bot inside hangs at the depth it has reached; when it opens a door, it drops another room. Click one to follow it through the maze, room by room.

    In the net

    Most recent catches · tagged on capture
    Pleroma
    GET /trap/ · 08:39:13
    ×24
    Masked visitor
    GET /trap/ · 19:05:53
    ×549
    Claude-User
    GET /trap/ · 15:12:24
    ×11
    Claude-User
    GET /trap · 15:12:24
    ×11

    Specimens in disguise

    Each of these arrived announcing itself as a well-known crawler. Real Googlebot comes from Google's network, and its address resolves to a google.com or googlebot.com hostname. These did not. (We don't keep visitors' addresses, so here is the network they came from instead.)

    Claimed to beActually came fromVerdict
    SofyaBot · 5 min agoAS203466 INTROSERV d.o.o · DEIMPOSTOR
    SofyaBot · 5 min agoAS203466 INTROSERV d.o.o · DEIMPOSTOR
    PetalBot · 14 min agoAS136907 HUAWEI INTERNATIONAL PTE. LTD. · SGIMPOSTOR
    PetalBot · 31 min agoAS136907 HUAWEI INTERNATIONAL PTE. LTD. · SGIMPOSTOR
    PetalBot · 52 min agoAS136907 HUAWEI INTERNATIONAL PTE. LTD. · SGIMPOSTOR
    PetalBot · 2 h agoAS136907 HUAWEI INTERNATIONAL PTE. LTD. · SGIMPOSTOR

    Escaped & recaptured

    The trap does not hold anything. It only remembers. These keep coming back.

    I.
    Masked visitor×549
    II.
    scraperbot9000×253
    III.
    Mastodon×240
    IV.
    curl command×69
    V.
    YaK×47
    VI.
    Unclassified automaton×40

    The Labyrinth

    Behind the staff door, a staircase goes down into a maze with no end. Every room has five doors, and every door leads deeper. Only bots that ignore robots.txt ever find it, and none has found the bottom, because there isn't one.

    11,672,691Rooms explored by bots
    110Species lost inside
    221,000Deepest room reached
    Deepest divers
    I.
    curl command949,967 rooms explored221000 deep
    II.
    MotherLoverBot47,053 rooms explored170737 deep
    III.
    Masked visitor5,501,505 rooms explored124414 deep
    IV.
    Slackbot124,927 rooms explored124382 deep
    V.
    scraperbot900023,083 rooms explored19645 deep
    VI.
    MySimpleCmdBot5,412 rooms explored4940 deep
    VII.
    KatherineBot3,996 rooms explored3780 deep
    VIII.
    kitbot1,778 rooms explored1812 deep
    IX.
    Googlebot (impostor)664 rooms explored1565 deep
    X.
    Perplexity-User (impostor)24 rooms explored1194 deep

    Follow a bot through the Labyrinth, live →

    The keepers stayed up watching curl command walk 949,967 rooms. They could use a coffee.

    Quick-change artists

    The only way down is one door at a time. So when a bot opens a door, deep in the Labyrinth, out of a room a different species has just left, it is the same walker in a new costume: a scraper that keeps changing its name.

    archive.org_bot
    ChatGPT-User
    curl command
    facebookexternalhit
    Masked visitor
    20 costume changes · last 38 days ago48,150 deep
    Set your own trap

    Add a disallowed path to your robots.txt, link to it invisibly, and log whatever arrives. Anything in that log ignored your rules.

    User-agent: *
    Disallow: /trap/