DropHunter

Your wishlist, quietly watched.

A private game and watch price tracker with a web dashboard, conversational Discord bot, and scheduled deal alerts. Track PC games across stores supported by IsThereAnyDeal, set a target in INR, and get notified when the price qualifies.

Quick start · Dashboard setup · Tracker operations · Issues

Actual application screenshot using the isolated 100-game demo. Prices, accounts, and counts are sample data.

What you can do

• Manage your watchlist: search, filter, sort, and paginate games; add a confirmed PC match, edit its target, or remove it. The web app permits up to 100 tracked games per user.
• Explore prices: select a game for its 7-, 30-, or 90-day history. See the store and timestamp behind each observation, with stale prices and history gaps shown explicitly.
• Get deal alerts: scheduled checks send Discord DMs, with optional per-user email. Use a custom target or the provider's historical low. Alerts still send when AI commentary is unavailable.
• Use natural language: manage games and Swiss Time House watches through an approved Discord DM conversation, with per-user memory.
• Keep access private: Google sign-in requires owner approval and an explicit link to the user's Discord watchlist. Users only access their own records.
• Run where it fits: game checks run on GitHub-hosted or self-hosted runners; the dashboard and bot deploy independently.

Screenshots

Google sign-in and account approval

The screen shows the current unconfigured local state. Live sign-in requires Google OAuth credentials, Supabase provider setup, and an approved account mapping. The demo needs none of these.

Quick start

Explore the dashboard without credentials

Requires Python 3.11 and Node.js 22. From the repository root:

python3.11 -m venv .venv
. .venv/bin/activate
pip install -r requirements-web.lock
npm --prefix frontend ci
npm --prefix frontend run build
flask --app 'web.app:create_app()' run --port 8081

Open http://127.0.0.1:8081 and choose Explore with sample data. Demo edits stay in that browser session and reset when the page reloads.

Connect live data

Follow dashboard setup for server configuration, the additive database migration, Google OAuth, account approval/revocation, and production hosting.

For the conversational bot and GitHub Actions checks, follow tracker operations. That guide includes required secrets, database setup, commands, email delivery, and home-server recovery.

Deployment at a glance

Component	Where it runs	Important dependency
Game price checks	GitHub Actions; cloud runner by default	PostgreSQL, ITAD key, Discord bot token
Web dashboard	Separate web service; cloud container or home server	Supabase Google Auth and approved account mappings for live access
Conversational bot	Persistent server	Discord connection and configured AI provider
Watch price checks	Self-hosted runner by default; opt-in	Runner that can reach the watch store
Database backup	Optional self-hosted workflow	Explicitly enable after reconciling cloud/home data

Game checks are scheduled every 12 hours at 00:00 and 12:00 UTC; GitHub can delay scheduled runs. GitHub Actions does not host the interactive dashboard or keep the conversational bot online.

Current scope: PC games are supported. PS5 tracking awaits a suitable pricing provider. The web dashboard lists watches; watch creation and conversational chat remain in the bot. Live dashboard activation is tracked separately from the implemented UI.

How it fits together

flowchart LR
    Browser[Web dashboard] --> Auth[Supabase Google Auth]
    Browser --> API[Python web API]
    API --> Approval[Owner-approved account mapping]
    API --> DB[(PostgreSQL)]
    Discord[Discord DMs] --> Bot[Conversational bot]
    Bot --> DB
    Actions[GitHub Actions] --> Checks[Price checks]
    Checks --> Prices[ITAD / watch store]
    Checks --> DB
    Checks --> Alerts[Discord + optional email]

Loading

The web API verifies the Supabase identity and checks approval before querying user-scoped data. Database credentials stay server-side. Existing Discord ownership is preserved. Use Supabase's public schema or the home server's drophunter schema.

Charts show the best recorded store price at each check, not a single store's complete history. The selected store can change, and gaps are not filled with invented prices.

Development and checks

With the virtual environment activated:

pip install -r requirements.txt -r requirements-dev.txt -r requirements-web.txt
python -m pytest -q
ruff check .
npm --prefix frontend ci
npm --prefix frontend test
npm --prefix frontend run build
# First browser-test run:
cd frontend
npx playwright install chromium
npm run test:e2e

CI runs Python tests against disposable PostgreSQL 17 databases in both supported schemas, plus frontend unit tests, the production build, and browser tests. Set TEST_DATABASE_URL to a disposable database to run database tests locally; otherwise those tests are skipped. Never use a production database for tests.

Repository map

Path	Purpose
frontend/	Vite dashboard, Supabase sign-in, browser tests
web/	Flask API, scoped queries, owner approval CLI
bot/ · ai/	Discord commands, LangGraph conversations, AI providers
cron/ · utils/	Scheduled checks, price sources, notifications
db/ · supabase/migrations/	Database clients, schemas, additive migrations
tests/	Python and PostgreSQL integration tests
Dockerfile.web · Dockerfile	Independent web and bot images

Documentation

• Dashboard setup, Google sign-in, approvals, deployment, and rollback
• Discord commands, cloud scheduling, notifications, and home-server recovery
• Dashboard design and security requirements
• Open issues and planned improvements