SeoSuss

AI-powered site structure mapping tool for SEOs based on target keywords.

Comment

Inspiration

As an SEO specialist, I’ve worked on numerous small to mid-sized campaigns. One of the most time-consuming parts of the job is developing a Keyword Strategy after the Keyword Research phase. This step involves:

  • Analyzing the SERPs (Search Engine Results Pages) for each keyword.
  • Comparing and grouping similar keywords into clusters.
  • Reviewing the client’s website to map those keyword groups to existing pages.
  • Identifying where new pages are needed.

Since I’ve always found keyword strategy formulation to be slow and tedious, I wanted to streamline the process. That’s what led me to build a tool to make it easier and more efficient.

What It Does

SeoSuss simplifies two key parts of SEO strategy:

  1. Keyword Clustering using two methods:

    • AI-Based Clustering: Groups keywords based on their semantic relevance.
    • SERP-Based Clustering: Groups keywords by comparing actual search results (SERPs).

  2. Site Structure Mapping based on finalized keyword clusters:

    • The tool crawls the entire website.
    • It recommends which existing pages should target each keyword group.
    • If no suitable page is found (based on a threshold), it suggests creating new ones.

How It Was Built

I built SeoSuss using Bolt.new, developing it incrementally:

  1. Keyword Clustering Module
  2. Website Crawling Module
  3. Site Structure Mapping
  4. Project-based architecture with persistent storage and user accounts.

Though it's still an MVP, it’s functional and the KW clustering part is already helping me in real projects.

Keyword Clustering Methods

  • AI-Based Clustering: Uses OpenAI’s text embeddings and cosine similarity to assess semantic relationships between keywords.
  • SERP-Based Clustering: Utilizes the DataForSEO API to fetch real-time SERP data and groups keywords based on overlap in search results.

Website Crawling

Initially, I considered using DataForSEO but found it didn’t support real-time crawling. After extensive research, I chose Spider.Cloud — a high-performance, cost-effective real-time crawler that fits perfectly with my needs.

Site Structure Mapping

Uses OpenAI’s text embeddings and cosine similarity — this time comparing the content (up to 1,000 characters) of each crawled page against the keyword clusters to find the best matches.

Challenges Faced

  • Integrating DataForSEO API: This involved a steep learning curve. I had to read all the documentation and handle request/response structures manually — especially challenging when batching multiple keyword tasks (since the API only accepts one task per request).

  • Interactive Sitemap Visualization: Designing an intuitive, interactive sitemap during the Site Structure Mapping phase took several iterations and brainstorming sessions with Bolt to get right (still improving).

Achievements I’m Proud Of

  • Despite being an MVP, building this tool during weekends and spare evenings is a big accomplishment. I see real potential in it and am optimistic about its future.

  • The Keyword Clustering feature, in particular, stands out. It took time to implement, but I now use it regularly and it’s made my work much more efficient.

Lessons Learned

  • I learned a lot about building applications with AI. While I had some coding experience before, creating this tool felt like something far beyond my skill set — until I did it. AI really reignited my old passion for building and experimenting.
  • The discussion mode in Bolt has been a game-changer. It feels empowering to build something while actively conversing with a tool that has access to vast knowledge. Now, I genuinely believe I can build anything if I dedicate time to it.
  • Always implement logging early — one major takeaway was realizing how important it is to log everything from the start. Having an application-wide logger helps catch bugs, especially those that fail silently and are otherwise hard to trace.

What’s Next for SeoSuss

Core Enhancements

  • Proper backend setup with user account management and settings.
  • Credit/token system to track usage per user (aiming for a pay-per-use model instead of subscriptions, as most SEOs only use such tools occasionally).

Planned Features

Keyword Clustering

  • Allow users to manually move keywords between clusters if needed.
  • Enable users to view and compare SERP results for each keyword when using SERP-based clustering.

Website Analysis

  • Stream crawl results in real-time to improve UX (users see results populate even before the crawl completes).

Site Structure Mapping

  • Allow manual reorganization of keyword clusters across pages.
  • Enable drag-and-drop reordering of site structure nodes.
  • Add AI-generated suggestions for URLs and page titles for newly proposed pages — tailored to the existing site structure.
  • Increase the amount of page content analyzed for better semantic matching with keyword clusters.

Built With

Share this project:

Updates