MetaTitle Tool

The MetaTitle Tool is a high-throughput meta title optimization system that operates entirely in vector space. For each URL, the pipeline: • Scrapes the page content and retrieves the URL’s highest-impact query from the Google Search Console API, using multi-parameter filtering (CTR, impressions, position, and relevance). • Generates semantic embeddings for: – the page content – the existing <title> tag – the top GSC query These embeddings are used to compute the semantic centroid and evaluate alignment between content and metadata. Because modern search engines rank pages using vector similarity rather than keyword matching, the system effectively interfaces with Google’s ranking logic by “speaking” in vectors. The tool then executes a generative candidate-selection loop: – OpenAI produces a batch of candidate meta titles – Each candidate is embedded – Cosine similarity is computed between each candidate vector and the content vector – The candidate with the highest semantic alignment is selected as the optimized meta title The workflow is built to run at scale, enabling bulk optimization of large URL sets with deterministic vector-based scoring rather than heuristic keyword methods. The result is a fully automated, semantically grounded meta title generator that maximizes alignment with search intent and search-engine vector semantics.

This project was built to help content creators and developers save time while improving website visibility. The engine uses vector embeddings to identify core themes from raw text, rewrite titles based on search intent, and ensure proper length, clarity, and keyword balance.