Docs

Embedding Recommendations

Hokai uses embeddings and semantic search to power Smart Match and directory search. This article explains how it works at a high level.

What Are Embeddings?

Embeddings are vector representations of text. Similar concepts map to nearby vectors. "AI coding assistant" and "developer copilot" are close in embedding space; "AI coding assistant" and "social media scheduler" are farther apart.

How Hokai Uses Them

Validation

Recommendations are validated against our tools database. We do not hallucinate tools. Every recommended tool exists in our directory with real data.

The Bottom Line

Embeddings help us find relevant tools. Validation ensures they're real. The result is a ranked stack that fits your context.

Related Docs