Text Embeddings

Embedding models convert text into numbers that capture meaning — the foundation for smart search, “find similar,” and retrieval features inside the tools your team builds.

Request early access Read the docs

What it’s good at

Meaning-aware search
Great for retrieval
Fast and inexpensive

Popular uses

Semantic search over documents
Recommendations
Powering AI assistants with your own data

How your team uses Text Embeddings

It’s available to everyone on your Roar account the moment you sign up — no separate subscription, no waiting. Non-technical teammates use it through the tools they build; developers call it directly by its model name bge-m3. One published price, and failover handled for you.

Other models