LMNFT New Collections Tracker

LaunchMyNFT is a platform where creators drop new NFT collections on a rolling basis. For a trader or a researcher, the interesting signal is the first few minutes after a collection goes live — mint price, supply, and how fast it is moving. The problem is that LaunchMyNFT does not push notifications when a new collection appears. You either refresh the page by hand or you miss it.

The problem

We needed to know about new NFT collections on LaunchMyNFT the moment they appeared, and we needed a persistent record of every collection and its mint activity over time. Manual monitoring does not scale — drops can happen at any hour, and the data that matters (mint progress, price changes, sell-through rate) changes minute to minute. By the time a human notices, the window is often closed.

The requirements were concrete: poll LaunchMyNFT for new collections on a tight interval, store every collection with its metadata and mint state, and push a Telegram alert the instant a new collection is detected. The alert had to carry enough information — name, supply, price, mint link — that a subscriber could decide whether to act without opening a browser.

The approach

We built a Telegram bot on aiogram paired with a polling script that does the heavy lifting. The stack: Python, aiogram for the bot framework, requests for HTTP, sqlite3 for storage. No external database, no message broker, no cloud functions. The whole thing runs as a single long-lived process.

The polling loop hits LaunchMyNFT's public collection listing on a fixed interval. On every response, it diffs the result against the known set of collections in the SQLite database. New collections — any collection ID not already in the collections table — trigger two things: a row insert with the full metadata, and a Telegram message broadcast to every subscribed chat.

The bot layer handles subscriber management. Users start the bot, pick which alert categories they care about, and the bot records their chat ID. When a new collection is detected, the script iterates over subscribed chat IDs and sends the alert. aiogram handles the Telegram API, rate limits, and retries — we do not reimplement any of that.

Mint activity tracking runs on a second schedule. For every collection in the database, the script periodically re-fetches its mint state — how many have been minted, current price, whether mint is still open — and appends a row to a mint_events table. This produces a time series per collection, which is what you need to compute sell-through rate and spot momentum.

What was built

A Telegram bot at @glitch_LMNFT_tracker_bot that any user can start and subscribe to. The bot exposes a small command surface: subscribe, unsubscribe, list recent collections, and query a specific collection's mint history. The subscription model is per-chat, so the same bot can serve individual traders and group chats with different alert preferences.

The backend is two loops sharing one SQLite database. The first loop polls for new collections and fires alerts. The second loop polls mint state for known collections and records the time series. SQLite was the right call: the write volume is low (a few rows per minute at peak), the read pattern is simple (latest state per collection), and the database is trivially backupable — it is a single file.

The alert payload is a formatted Telegram message with the collection name, cover image, supply, mint price, and a direct mint link. It is designed to be scannable in a notification preview — the name and price fit on one line — so a subscriber can triage without opening the chat.

Results

• Automated NFT collection tracking with new drops detected within one polling interval of going live.
• Telegram alerts pushed to every subscribed chat the moment a new collection appears, with enough metadata to act on.
• Mint analytics as a per-collection time series, enabling sell-through rate and momentum calculations after the fact.

The bot is live and serving subscribers. The SQLite database holds the full history of every collection the tracker has seen since launch, which is the real asset — the alerts are ephemeral, but the time series is what you mine later for patterns.

The takeaway

The value of a tracker is not the alert — it is the record. Alerts get acted on or ignored in the moment, but the persistent store of every collection and its mint trajectory is what lets you look back and ask "what did the winners have in common?" Build the alert to be fast and the storage to be durable. The alert is the product; the database is the moat.