Summary

This project extracts actionable investment insights from social media by tracking an influencer's posts. It employs AI for contrarian analysis, identifies assets, predicts market movements, and delivers real-time, multi-platform notifications, offering users automated, AI-driven trading signals.

How It Works

The system scrapes Facebook posts via Apify, using LLMs (e.g., Claude Opus) for analysis and contrarian logic ("reverse deduction") to infer market shifts. It automatically maps identified assets, records baseline prices, and tracks actual price movements (OHLC) for five trading days. Chain-of-thought reasoning captures indirect effects. Data is stored in SQLite, with results pushed via Telegram, Discord, or LINE.

Quick Start & Requirements

Deployment options include one-click Zeabur, Docker (docker build, docker run), or npx (npx @cablate/banini-tracker serve) for Node.js users. Local development requires npm install && npm run start. Essential prerequisites are Apify and LLM API keys. Optional keys support notification services and stock data. Web UI configuration runs on port 3000.

Highlighted Details

• AI-driven contrarian analysis of social media influencer "Ba Ni Ni (8zz)".
• Automated 5-day prediction tracking with a "jinx success" metric (>1% move).
• Historical backtesting (2024-2026) on 2,249 posts, yielding 345 explicit predictions.
• Command-line interface (CLI) available via npx for direct use or integration.

Maintenance & Community

Maintained by cablate. No specific community channels or roadmap links are provided. The recent relicensing indicates ongoing development.

Licensing & Compatibility

Licensed under AGPL-3.0 (switched from MIT on 2026/04/13). This strong copyleft license mandates that modifications and network service deployments must also be released under AGPL-3.0. Compatibility with closed-source or commercial applications is restricted due to source-sharing obligations.

Limitations & Caveats

AGPL-3.0 imposes significant obligations for network service deployments. Reliance on external APIs (Apify, LLMs) introduces potential costs and rate limits. Effectiveness depends on the influencer's posting habits and AI interpretation accuracy. LINE notifications are capped at 200/month on the free tier.