CRWLR.OS 0.2.0 / LOCAL ENVIRONMENT SANDBOX ONLINE / NO EXTERNAL REQUESTS
⌁ CRAWLERS[ ALPHA ]
ENVIRONMENT_001 / SYNTHETIC WEBT+00000
AN OPEN EXPERIMENT IN AUTONOMOUS SYSTEMS

Your rules.
Their instinct.

Build a world. Release your crawlers.
See what happens when curiosity runs loose.

FIG. 001 — THE SWARMBEHAVIOR: EMERGENT
CONTAINMENT: LOCAL
OBSERVER: YOU
PROCESS FEED SIM
+
+
THE MACHINE IS CURIOUS. ARE YOU?GENERATIVE SYSTEM / NO. 001
UNTITLED_EXPERIMENTLIVE LOCAL SIMULATION
NODES0000
FRONTIER0001
NOT A SEARCH ENGINE.
A PLACE TO LET THINGS RUN.

Your mission. Your parameters. Your little corner of chaos.
Configure a simulation below, or fork an experiment and make it yours.

AVAILABLE ENVIRONMENTS06
EXECUTIONLocal/ YOUR DEVICE
NETWORK ACCESSIsolated/ SYNTHETIC
SOURCE CODEOpen/ MIT
SELECT AN ENVIRONMENT

Enter the sandbox_

FORK. MODIFY. REPEAT.
SHOWING 6 INDEPENDENT INTELLIGENCES[ EVERY RUN STARTS WITH A QUESTION ]
// THE PROTOCOL

Let something
off the leash.

01 — DEFINE

Write the directive.

Give the swarm a mission. Set its population, exploration depth, and node budget.

02 — RELEASE

Run the experiment.

Watch workers traverse a synthetic web. Change the speed. Pause the system. Follow every event.

03 — EVOLVE

Change the conditions.

Inspect the run, fork its configuration, and see how a different swarm behaves.

while (curiosity) { explore(); }

Define the
experiment.

Set the conditions. The swarm takes it from here.

◉ Simulation mode — your crawler is saved on this device. No websites are fetched and no model is trained.

Observe.
Interfere. Repeat.

Crawlers is an open-source environment for running your own crawler simulations. Set the conditions and observe the swarm.

Runs traverse a deterministic synthetic graph in your browser. Worker population controls concurrency; depth shapes the graph; the node budget bounds each run. Pause, resume, step, reset, change speed, or fork a configuration. Runs advance only while this page is open.

Mission text and starting URLs describe intent; they do not affect the synthetic graph. No sites are fetched. Talk is a scripted preview, not an LLM. Live web exploration and source-backed intelligence are planned for a later stage.

No account. No API keys. Local browser storage only. Export your crawler configuration from its profile.