Crawlbet DEMO — the soccer intelligence network

~/crawlbet ❯ agents watch --soccer ▌

❯ autonomous agents.
the beautiful game. a sharper edge.

A swarm of crawl agents reads team news, form, lineups and market movement. Every match becomes a probability. Every signal comes with the reasoning.

4 agents awake / 6 matches under watch / 24/7 curiosity

❯less noise. more signal.
watch the swarm find its next edge.

[ built for the edge. driven by the game. ]v0.1

xg-weaver / watching the pitch

ENG · PREMIER LEAGUEsample fixture

ArsenalvsChelsea

Emirates Stadiumpre-match analysis ↗

illustrative analysis

home pressure away pressure

▰▰▰▰▰▰▰▱▱▱ 72% analysis complete128 data points

xg-weaver›comparing home xG against market probability…

SWARM SIGNALArsenal · match winner

model61%

edge+6.1pp

SWARM STATUS 4 agents scanning

MARKETS MONITORED18 across 3 leagues

SAMPLE PAPER RETURN+2.69u / 8 settled

NETWORK MODEdemo / no real wagers

01 / the swarm

CPU |||||||||||||||||| 68%
INT ||||||||||||| 4 / 4 awake

Tasks: 4 agents, 18 markets
Focus: soccer / pre-match

PID	AGENT	S	FOCUS	SCANS	EDGE

❯ select an agent to follow its analysis[4 running]

02 / activity streamsimulated feed

stdout ●streaming demo events

03 / signals worth a closer look

~/crawlbet ❯ swarm --list

meet the swarm_

Four specialist agents. Different lenses on the same game.

active agents / demo network

~/crawlbet ❯ signals --open

follow the signal_

See the market, the model and the reasoning behind every sample pick.

6 sample signalsdemo probabilities / decimal odds

~/crawlbet ❯ ledger --settled --paper

every result. on record_

A transparent sample ledger. The wins, the losses and the full picture.

PAPER PROFIT

WIN RATE

RETURN ON STAKE

SAMPLE SIZE8 picks

cumulative paper P/L / fixed 1u stakeillustrative results

MATCH	MARKET	ODDS	RESULT	NET P/L

Demo ledger using illustrative fixtures and outcomes. 1u is a fictional stake unit. Paper results do not establish future performance.

~/crawlbet ❯ man crawlbet

the field manual_

Understand what the swarm sees. Make up your own mind.

01 / crawl

Read the game.

Agents inspect the signals that shape a match: recent form, expected goals, team news, lineup availability and market movement. Each agent focuses on one part of the picture.

02 / compare

Price the probability.

A model probability is an estimate, not a promise. Market implied probability is 1 ÷ decimal odds. Edge is the difference between the estimate and that market probability.

61% − (1 ÷ 1.82) = +6.1pp

03 / track

Keep the whole ledger.

Inspect the reasoning before saving a signal. Track every result, including losses. A positive edge can still lose; a short winning run is not proof of a profitable model.

What is this version of Crawlbet?

An interactive concept website with a simulated swarm, illustrative soccer signals and a paper ledger. Match analysis, odds and results are sample data. This version does not connect to bookmakers, read live feeds, run AI models or place bets.

What does spawning an agent do?

It creates a local agent profile in your browser. Choose a name, league and focus. Your profile stays in this browser and can be removed from the agents screen. It does not deploy a server or execute real analysis.

How does the watchlist work?

Save sample signals with the star button. Your selections persist on this device, and the watchlist filter shows only saved signals. No sign-in is needed.

What are confidence, edge and units?

Model probability estimates how often an outcome might happen. Edge is measured in percentage points (pp) against unadjusted market probability. A unit (u) is a fictional stake for paper accounting. None of these measures guarantees a result.

crawlbet — soccer intelligence, in motion.

meet the swarm_

follow the signal_

every result. on record_

the field manual_

Read the game.

Price the probability.

Keep the whole ledger.