manzibonfils09-creator/afrisignal

AI-powered financial intelligence newsletter for African markets — built on QuantMind + Claude + GPT-4o-mini # AfriSignal 🌍 ### AI-Powered Financial Intelligence for African Markets > **Weekly digest of African fintech, crypto regulation, macro signals — extracted by AI, curated for practitioners.** ## What Is AfriSignal? AfriSignal is an automated intelligence pipeline that: 1. **Ingests** financial news, regulatory updates, and macro reports from African sources daily 2. **Extracts** structured signals using GPT-4o-mini via QuantMind's pipeline 3. **Synthesizes** a weekly digest using Claude Sonnet 4 4. **Publishes** via email newsletter (Resend), web dashboard, and Telegram channel Built on top of QuantMind — the open-source quant knowledge extraction framework accepted at NeurIPS 2025. ## Coverage Universe | Category | Sources | |----------|---------| | **Macro** | AfDB Reports, IMF Africa Outlook, World Bank, EAC Bulletins | | **Crypto/Fintech** | CoinDesk Africa, Techpoint Africa, Disrupt Africa, Blockchain Africa | | **Rwanda** | BNR (National Bank), RDB updates, New Times | | **Kenya** | CBK regulatory notices, Business Daily, TechCabal | | **Nigeria** | CBN circulars, Nairametrics, Techpoint | | **Research** | arXiv (Africa-relevant quant papers) | ## Architecture ``` Scheduler (daily 06:00 UTC) → Ingestion Layer (RSS, Web Scrape, PDF, arXiv API) → QuantMind Processing (signal extraction via GPT-4o-mini) → Knowledge Store (PostgreSQL + pgvector) → Weekly Digest (Claude Sonnet synthesis, Monday 07:00 UTC) → Delivery: Email (Resend) + Web Dashboard (FastAPI) + Telegram ``` ## Quick Start ```bash git clone github.com cd afrisignal curl -LsSf astral.sh | sh uv venv && source .venv/bin/activate uv pip install -e . cp .env.example .env # Fill in your API keys, then: python -m afrisignal.cli ingest --date today python -m afrisignal.cli digest --week current --dry-run ``` ## Cost: ~$18/month at launch See docs/COST_BREAKDOWN.md for full breakdown. See docs/MONETIZATION.md for revenue model …