Narratex

Crypto narrative intelligence powered by Binance Square.

Narratex monitors Binance Square in real time and detects which crypto narratives are gaining momentum — before the crowd arrives. Instead of tracking individual tokens, it tracks the themes driving the market.

🔗 Dashboard: https://davexinoh.github.io/Narratex/dashboard.html 🤖 Telegram Bot: https://t.me/Narratexbot

---

What It Does

Most tools track prices. Narratex tracks narratives.

It scans Binance Square posts, scores them against 8 narrative clusters, and surfaces which themes are accelerating — along with the tokens most likely to benefit.

Example output:

AI Infrastructure     ████████ 84%  ↑ RISING
Tokens: FET · TAO · RNDR · AKT
Mentions +91% · Engagement +88%

Solana Ecosystem      ███████░ 79%  ▲ PEAK
Tokens: SOL · JUP · RAY · BONK

DePIN Compute         ██████░░ 67%  ↗ EMERGING
Tokens: HNT · IOTX · FIL · AKT

---

Architecture

Binance Square (public feed)
        ↓
collector.py      — fetches posts, scores against narrative seeds
        ↓
extractor.py      — clusters signals into named narrative buckets
        ↓
momentum.py       — weighted confidence scoring + lifecycle stage
        ↓
api.py (Flask)    — /api/narratives with 5-min cache
        ↓
Render            — always-on backend deployment
        ↓
dashboard.js      — fetches API, renders 9-section dashboard
        ↓
GitHub Pages      — static frontend

---

Confidence Score

Each narrative is scored using three signals, normalized to 0–100 and weighted:

Signal	Weight	Source
Mentions Growth	50%	Binance Square post frequency
Engagement Growth	30%	Likes + comments + shares
Volume Momentum	20%	Binance public ticker (24hr change)

---

Lifecycle Stages

Each narrative is assigned a stage based on confidence and signal velocity:

Stage	Meaning
↗ EMERGING	New narrative forming, low confidence but accelerating
↑ RISING	Gaining momentum, worth watching
▲ PEAK	Maximum narrative strength
↘ DECLINING	Losing momentum

---

Dashboard — 9 Pages

#	Page	Description
01	Narrative Leaderboard	All narratives ranked by confidence
02	Lifecycle Tracker	Narratives grouped by market cycle stage
03	Narrative Heatmap	Color-coded confidence grid
04	Signal Breakdown	Bar chart of all three signals per narrative
05	Token Radar	Top 10 tokens by narrative strength
06	Active Narratives	Full signal breakdown with token lists
07	Capital Rotation	Liquidity flow between narrative sectors
08	Narrative Token Map	Every narrative cluster with all associated tokens
09	Predictions	Emerging narratives scored by signal acceleration

---

Telegram Bot

@Narratexbot delivers live narrative intelligence on demand.

Command	Response
/start	Introduction and command list
/briefing	Full narrative intelligence briefing
/leaderboard	Top narratives ranked by confidence
/rotation	Capital flow between narrative sectors
/predictions	Predicted emerging narratives
/tokens <narrative>	Tokens associated with a specific narrative
/refresh	Force re-fetch of live Binance Square data

Or ask anything in plain text — the bot uses AI to answer based on live narrative data.

---

OpenClaw Skill

Narratex ships as an OpenClaw-compatible skill in skill/.

# Install
git clone https://github.com/Davexinoh/Narratex ~/.openclaw/skills/narratex

# Or via ClawHub
clawhub install narratex

Trigger phrases: narrative briefing, what's trending, token radar, what's emerging

---

Tech Stack

Layer	Technology
Data source	Binance Square public feed
Market data	Binance REST API (/api/v3/ticker/24hr)
Backend	Python 3.11, Flask, Gunicorn
Frontend	HTML, CSS, JavaScript, Chart.js 4.4.2
Bot	python-telegram-bot 21.5
Backend hosting	Render
Frontend hosting	GitHub Pages

---

API Endpoints

Method	Endpoint	Description
GET	/	Health check
GET	/api/narratives	All narratives sorted by confidence
GET	/api/narratives/<name>	Single narrative detail
GET	/api/status	Cache metadata

Query params on /api/narratives:

• ?refresh=true — bypass cache and re-run pipeline
• ?min_confidence=N — filter by minimum confidence score

---

Repo Structure

Narratex/
├── backend/
│   ├── api.py           — Flask API
│   ├── collector.py     — Binance Square signal collector
│   ├── extractor.py     — Narrative clustering engine
│   ├── momentum.py      — Confidence scoring + lifecycle stage
│   └── seeds.py         — Single source of truth for narrative seeds
├── data/
│   └── narratives.json  — Seed/fallback data
├── skill/
│   ├── SKILL.md         — OpenClaw skill definition
│   └── scripts/
│       ├── fetch_narratives.py
│       └── briefing.py
├── bot.py               — Telegram bot (runs on Render)
├── dashboard.html       — 9-page dashboard
├── dashboard.js         — Data fetching and rendering
├── index.html           — Landing page
├── styles.css           — All styles
├── render.yaml          — Render deployment config
└── requirements.txt

---

Deployment

Backend (Render):

• Auto-deploys from main branch
• Single Gunicorn worker (required for in-memory cache)
• TELEGRAM_TOKEN and ANTHROPIC_API_KEY set as env vars

Frontend (GitHub Pages):

• Static files served from repo root
• No build step required

---

Built for the Binance OpenClaw AI Hackathon 2026 · by Davexinoh and Xavier