Inspiration
Admins of crypto and trading communities spend hours every day manually tracking asset prices, generating charts, and writing technical analysis updates.
• This repetitive work doesn't scale.
• It often leads to burnout, inconsistent updates, and a drop in community engagement.
We wanted to build a tool that completely automates this process—providing institutional-grade insights on autopilot so community leaders can focus on actually
engaging with their members instead of staring at charts.
## What it does
Re-Stocker is a Multi-Indicator Quantitative Signal Scanner for social communities.
• Monitors: It continuously scans a customizable watchlist of assets (crypto, stocks, forex).
• Calculates: It runs the math on core technical indicators, including SMA20, Bollinger Bands, RSI-14, and MACD.
• Scores: Using a composite scoring engine, it filters out market noise and evaluates the absolute strength of a trend.
• Alerts: When it detects a high-confidence opportunity (either bullish or bearish), it automatically broadcasts a rich alert with a custom multi-panel chart
directly to a Discord channel.
## How we built it
• Language & Data: We built the core logic using Python. We used yfinance to fetch real-time market data.
• Math Engine: We used pandas alongside numpy to calculate all technical indicators from scratch. By writing the math ourselves instead of relying on heavy, C-
dependent TA libraries, we kept the bot lightweight and highly portable.
• Visuals: The dark-mode, multi-panel charts were rendered using matplotlib.
• Automation & CI/CD: We set up a CI/CD pipeline using GitHub Actions, leveraging cron schedules to run the bot autonomously every hour.
• Integration: The bot uses the requests library to securely format and send rich embedded messages to a Discord Webhook.
## Challenges we ran into
• Signal Noise: A significant challenge was creating a robust composite sentiment engine that didn't just spam the community channel. We had to heavily fine-tune
our indicator logic and introduce an alert threshold system to ensure the bot only triggered on genuine, high-confidence trading opportunities.
• Headless Constraints: Ensuring our mathematical operations and chart generation worked flawlessly within the constraints of a headless GitHub Actions runner
required careful dependency management.
## Accomplishments that we're proud of
• Professional Charting: We are incredibly proud of the multi-panel dark-mode charts—they look highly professional and mimic the quality you would expect from
premium platforms like TradingView.
• Serverless Architecture: We're also proud of our "set it and forget it" architecture. By leveraging GitHub Actions as a scheduling backend, we deployed a
completely serverless, autonomous bot that operates 24/7 at zero hosting cost.
## What we learned
• Quantitative Analysis: We learned a ton about quantitative indicator calculations and how to manipulate financial time-series data using pure Pandas and Numpy.
• Pipelines & Webhooks: We gained valuable experience designing serverless automation pipelines and working with Discord's webhook API to format complex,
aesthetically pleasing embedded messages.
## What's next for Re-Stocker
• Multi-Timeframe Analysis: Cross-referencing 1H, 4H, and 1D charts to identify stronger, aligned signals. • Broader Reach: Expanding our communication channels to include Telegram bots and daily email digests. • Web Dashboard: Building a dedicated web application to track the historical accuracy and performance of the bot's signals. • Volume Analysis: Adding volume-based breakout detection to our composite scoring engine for even higher-fidelity alerts!***
