SaaS Map: a Tool Catalog with AI-Automated Weekly Pricing Updates

One of the most annoying parts of starting a new project is comparing SaaS tools: hosting, databases, authentication, transactional email… Every provider has its own pricing page, often in a different language, with plans that change every couple of months.

I built SaaS Map to solve exactly that: a catalog of 30+ tools with filters, a side-by-side comparator, and pricing that updates itself every week.

The Problem I Wanted to Solve

Every time I start a new project I end up opening ten browser tabs to answer basic questions:

• Which serverless database has a generous free tier?
• Which hosting platform gets cheaper at higher usage?
• Which email service has the best deliverability for LATAM?

The information exists, but it’s scattered, often in English only, and frequently outdated. SaaS Map brings it all into one place — in Spanish (and also EN, PT, and FR) — with up to 3 tools compared side by side.

The Interesting Part: the Auto-Update Pipeline

What makes this project different from a static directory is that both the tools and the news section update automatically. The pipeline works like this:

Tool Updates

1. A crawler visits each tool’s official website — specifically the pricing page and homepage.
2. It extracts the relevant content: free plan details, base price, and key features.
3. That raw data is passed to an AI model which summarizes, normalizes, and translates it into the four site languages.
4. The result goes straight into the database, with a last-updated timestamp visible on each tool card.

This means when Resend drops their price or Supabase changes their free tier limits, the next crawler run reflects it with zero manual intervention.

Automatic News Generation

The same pipeline feeds the site’s news section. When the crawler detects significant changes in a tool — a new plan, a price change, a new feature — the AI generates a short editorial note that appears on the site’s blog.

It’s not just an external RSS feed: it’s content generated from real changes detected in the catalog tools.

The Stack

• Astro for the static site with partial hydration
• Astro’s native i18n for the four languages
• The crawler runs on a self-hosted server with a weekly cron job
• AI (Claude or GPT-4o mini depending on the task) processes the text and generates the news
• Relational database to store the price change history

What I Learned

Prices change more often than you’d expect. In the first month of operation, the crawler detected changes in 4 out of 30 tools. Without automation, that would have been hours of manual work.

Cheap LLMs are good enough for summarization tasks. I don’t need GPT-4o to summarize a SaaS pricing page. GPT-4o mini or Claude Haiku handle it perfectly at a fraction of the cost.

Automatic translation has limits. For short technical texts it works great. For longer copy with heavy business context, manual review is still needed.

The free tier matters more than vendors want to admit. The “free plan only” filter is the most-used on the site. Developers in LATAM care about getting started without a credit card.

What’s Next

• Expand the catalog to 50+ tools with new categories (monitoring, analytics, payments)
• Email alerts when a tool you bookmarked changes its pricing
• Price history chart per tool

If you’re putting together a stack and want to compare options before committing, give it a try.