YakWP AI chatbot for WordPress — feature overview

YakWP Changelog — What’s New in Every Release

Here is what’s new in every YakWP release. Updates are delivered automatically to your WordPress dashboard under Plugins → Updates, and a fresh zip is always available on the Download page.

2.4.0 — Product-aware semantic search

  • Products are now indexed with their price, SKU, and categories. The chatbot can quote accurate prices straight from your WooCommerce store instead of guessing.
  • Scales to large catalogs. Retrieval is now bounded, so answers stay fast even when your store has thousands of products.
  • More reliable automatic indexing. The daily index now self-schedules on every page load, fixing a rare issue where in-place upgrades skipped automatic daily indexing.

2.3.4 — Semantic search on the free plan

  • Semantic (meaning-based) search is now fully available on the free plan — no Pro license required to build the index.
  • Free installs can now answer by meaning (e.g. “how much does it cost?”), not just exact keywords.

2.3.2 — Smarter default quick replies

  • Better default Quick Reply buttons — natural visitor-phrased questions that engage visitors and pair well with semantic search.
  • Auto-upgrade for sites still on the old default replies (custom replies are preserved).

2.3.1 — Better identity handling

  • The default system prompt now handles identity questions (who you are, where you’re located, are you human) and points visitors to the site’s contact page for contact or location queries.

2.3.0 — Semantic search is now Free

  • Semantic (vector) search is now included on the free plan — matches content by meaning, not just keywords.
  • Free sites show the “Powered by Yak” footer link; Pro can white-label.

2.2.1 — Faster, better retrieval

  • Switched to a modern embedding model (gemini-embedding-001) with 768-dimension vectors and task-aware retrieval for higher-quality matches.
  • Added a daily scheduled index batch (up to 1,000 pages/day) so large sites build their index without hitting rate limits.

2.2.0 — Semantic search + one-click indexing

  • Semantic (vector) search understands natural-language questions by meaning (e.g. “how much does it cost?” matches “$29”).
  • Built-in embedding index with one-click build/rebuild and auto re-indexing on save.
  • Refreshed the OpenRouter free-model list (removed dead models, added live ones).

2.1.7 — OpenCode Go provider

  • Added the OpenCode Go provider with 19 models, including its low-cost first-month plan, plus automatic base-URL routing.

Questions about a release? Contact us or check the in-plugin changelog in your dashboard.