Expand description
Provider-agnostic embedding abstractions.
Embeddings are numerical representations of text or other inputs. Rig uses
an embedding Model to generate vectors, Embed to
select which text from a Rust value should be embedded, and
EmbeddingsBuilder to batch embedding requests for vector stores or
retrieval workflows.
use rig_core::embeddings::to_texts;
assert_eq!(to_texts(vec!["first", "second"])?, vec!["first", "second"]);Re-exports§
pub use builder::EmbeddingsBuilder;pub use embed::Embed;pub use embed::EmbedError;pub use embed::TextEmbedder;pub use embed::to_texts;pub use tool::ToolSchema;pub use embedding::*;
Modules§
- builder
- Batched embedding generation for documents implementing
Embed. - distance
- Similarity and distance of embedding vectors.
- embed
- Text extraction for embedding builders, including implementations for common types.
- embedding
- Text and image embedding models, responses, and input identifiers.
- tool
- Embeddable tool descriptions and serialized tool context.