How I Track AI API Prices in Real-Time Using a $5 Alibaba Cloud Server

Description and use

Built a real‑time AI‑API price tracker on a $5 VPS that scrapes pricing data, stores it in SQLite, visualizes trends, and alerts via Telegram when prices drop.

Problem

Developers need up-to-date AI API pricing to avoid overpaying, but official pages change frequently and lack automation.

Outcome

The setup runs with <5% CPU, costs <$6/year, and has already saved the author hundreds of dollars by enabling dynamic routing to cheaper providers.

Workflow

• 1. Write a Python script using requests and BeautifulSoup to scrape public pricing pages or call machine-readable APIs.
• 2. Store fetched price data in a local SQLite database.
• 3. Schedule the script with Linux cron to run every 6 hours.
• 4. Serve a static HTML/JS dashboard via Nginx that reads the SQLite DB and renders Chart.js graphs.
• 5. Configure a webhook to send a Telegram message when a tracked model’s price decreases by more than 5%.
• 6. Deploy the stack on a low-cost Alibaba Cloud Lightweight Server (2C2G/40G ESSD) for under $5.3/year.

Evidence

I use a Python script using requests and BeautifulSoup to parse public pricing pages, combined with direct API calls for providers that offer machine-readable rate cards.

Source

SQLite for local, zero-config storage of historical price data.

Source

Standard Linux cron running every 6 hours to fetch daily-verified data without triggering aggressive rate limits or IP bans.

Source

A lightweight static HTML and vanilla JavaScript dashboard served via Nginx, displaying interactive graphs for price trends over time.

Source