GitHub Repository

Automated financial market digest platform — fetches data, runs analysis, delivers via Telegram

3 starsPython

Market Digest: Self-hosted market analysis and Telegram

Name: Market Digest: Self-hosted market analysis and Telegram
Availability: InStock
Author: mutaaf

by mutaaf·Feb 27, 2026·1 point·2 comments

Visit Project View on HN

AI Analysis

●●SolidSolve My ProblemNiche Gem

Consolidates pre-market routine into Telegram digest with backtestable scoring.

Strengths

•Retrace system tracks your picks against actual next-day prices—actionable feedback loop
•84 pre-configured instruments (stocks, forex, crypto, commodities) with custom data source support
•Fully self-hosted, MIT licensed, no subscription—removes friction for retail traders

Weaknesses

•Scoring methodology not documented—unclear if 0-100 grades are arbitrary or validated
•No evidence the technical analysis accuracy is better than TradingView or other established platforms

Post Description

Hi HN, I built this because my pre-market routine was eating 45+ minutes every morning — checking TradingView, Finviz, economic calendars, news sites, fear/greed indexes, all before market open.

*What it does:* Market Digest pulls data from 6 free sources (yfinance, TwelveData, Finnhub, FRED, NewsAPI, Fear & Greed), runs multi-timeframe technical analysis (daily/weekly/monthly RSI, pivot points, trend detection), scores every instrument from 0-100, and sends you a formatted digest on Telegram.

*What makes it different:* - Fully self-hosted, MIT licensed, no subscription - 84 instruments pre-configured (stocks, forex, commodities, crypto, indices) add more as you need - Multi-timeframe scoring — day trade, swing, and long-term grades with separate weight configs - Retrace system — snapshots every pick and grades it against actual next-day prices, so you see your accuracy over time - Equity fundamentals scoring (valuation, profitability, growth, health) - Custom data sources — plug in any HTTP API, RSS feed, or CSV file through the web UI (no code) - Optional AI commentary via Claude/GPT/Gemini with 3-provider fallback - Web UI ("Command Center") with 11 pages for config, preview, scorecards, and performance tracking

*Tech stack:* Python 3.12, FastAPI (60 endpoints), React + TypeScript + Tailwind, yfinance for data, dual-tier cache (memory + JSON). No database — all persistence is YAML configs + JSON files.

*Zero API keys needed to start* — `./setup.sh` → `make digest-dry` gives you a working digest with yfinance data. The README includes a step-by-step Telegram setup guide (takes 2 minutes), and everything is explained for people who've never used a terminal.

Happy to answer questions about the scoring algorithm, architecture, or anything else.