A plain-English explanation of embeddings, the numerical fingerprints of meaning that let PurioChat understand what visitors mean rather than just matching keywords.
Training & Semantic Search
Train PurioChat on content, documents, external pages, custom fields, embeddings, and semantic search.
9 articles
Training Your Content (and When to Retrain)
How to run training from the Data Training tab, what the progress bar shows, and when you need to retrain your content.
Choosing Content Types to Train
A guide to which content types PurioChat can train on (Posts, Listings, Pages, WooCommerce Products, Custom Post Types), which are Free vs Pro, and exactly what each one indexes.
Training on Documents (PDF, TXT, MD, XML, CSV)
Upload PDF, TXT, MD, XML, and CSV files so the AI can answer from content that isn’t published on your website.
Training on External Pages
Add external URLs so PurioChat fetches their main content, indexes it instantly, and uses it to answer visitor questions.
Custom Fields for AI Training
Include extra custom fields (ACF, WooCommerce, theme/plugin meta) in AI training per post type, with optional AI auto-detection of useful fields.
Choosing an Embedding Model
Pick the embedding model and dimension count that converts your content into vectors for semantic search, and learn why changing it means retraining.
AI Semantic Search & the Search Field
How PurioChat’s AI semantic search understands meaning instead of keywords, and how to add a standalone search box with the [ai_search_field] shortcode.
Search Quality & Match Thresholds
Tune how strict PurioChat is about which content qualifies as a match, using the Minimum Match Percentage and Best Match Badge thresholds.
