Crate agentic_vision 

AgenticVision — core vision library for image capture, embedding, similarity, and visual memory.

Re-exports

pub use capture::capture_from_base64;

pub use capture::capture_from_file;

pub use capture::generate_thumbnail;

pub use diff::compute_diff;

pub use embedding::EmbeddingEngine;

pub use embedding::EMBEDDING_DIM;

pub use similarity::cosine_similarity;

pub use similarity::find_similar;

pub use storage::AvisReader;

pub use storage::AvisWriter;

pub use types::*;

Modules

capture

Image capture and thumbnail generation.

diff

Change detection between visual captures.

embedding

CLIP embedding generation via ONNX Runtime.

similarity

Vector similarity search for visual embeddings.

storage

.avis binary file format reader/writer for visual memory.

types

Core data types for visual observations and memory.