Build
03concept system

Narrative Engine

Crypto signal → tomorrow's content narrative

Crypto Twitter + market data + Polymarket → the narrative before it's consensus.

A signal engine that listens to crypto Twitter, cross-checks on-chain market data, and reads Polymarket odds — then synthesises the emerging narrative with a conviction score and a timing window, so content ships while the story is still forming, not after it's priced in.

3 signals
social · market · odds, correlated
conviction
scored — not just trending volume
48–72h
typical lead on consensus
Concept system — fictional signals, real architecture. Shown to demonstrate how I'd build the correlation layer.
— the old way

In crypto, narrative is the product — and by the time a trend is obvious on the timeline, the attention is gone. Reacting is losing. The hard part is seeing the narrative form across three noisy, unrelated signals at once.

— the workflow
01Listen
Apify · CT scrape
02Market
market data API
03Odds
Polymarket API
04Synthesise
Claude · conviction model
05Predict
scoring · timing model
06Brief
Claude · angle generator
— the system, running

Sentiment velocity, market data and prediction-market odds resolve into ranked narrative calls with content angles attached.

Narrative Enginerunning
crypto twittersentiment velocity
@kol_alpharestaking yields are quietly repricing
@onchain_liLRT inflows +3 days straight
@desk_notesnobody's talking about this yet
market data
ETH
$3,840+5.2%
EIGEN
$4.10+11.8%
BTC
$96.2K+1.1%
polymarketodds · YES
Restaking TVL > $25B by Q371%
Major LRT listing this month64%
narrative predictions · conviction-ranked

Restaking rotation

window 48–72h·3 content angles
78%
conviction

LRT inflows + Polymarket odds lead the timeline. Ship the explainer before it's consensus.

Stablecoin yield migration

window 4–6 days·2 content angles
61%
conviction

Building, not loud yet. Watch funding; queue the contrarian take.

— stage by stage
01Listen

Crypto Twitter velocity

Tracks KOL posts, mention volume and — more importantly — the rate of change of sentiment. A narrative accelerating from a low base matters more than one already loud.

Apify · CT scrape
02Market

Confirm on-chain

Pulls price, volume, funding and flows for the assets a narrative implicates — to separate real positioning from pure timeline noise.

market data API
03Odds

Read Polymarket

Reads prediction-market odds on related events. When the crowd is putting money on an outcome the timeline hasn't caught up to yet, that gap is the signal.

Polymarket API
04Synthesise

Correlate the three

Correlates social velocity, market confirmation and odds movement into candidate narratives — and scores each on conviction, not just volume.

Claude · conviction model
05Predict

Rank with a window

Ranks narratives by conviction and attaches a timing window — how long before this is consensus. The whole point is to be early, with a reason.

scoring · timing model
06Brief

Angles, attached

For each high-conviction call, drafts content angles — hooks, formats, the contrarian take — so the narrative arrives as a shippable brief, not a chart.

Claude · angle generator
— the system underneath

Sentiment tools exist. Market dashboards exist. Polymarket is public. The build is the correlation layer that fuses all three and outputs a conviction-scored call with a timing window and content angles attached — the difference between a dashboard you have to interpret and a brief you can act on. In an attention market, a 48-hour lead on the narrative is the entire edge.

— stack
Apifycrypto Twitter scrape
Market data APIprice / volume / funding
Polymarketprediction-market odds
Claudesynthesis + conviction scoring
signalcryptoPolymarketnarrative