semantic_recall

Overview ¶

Example: semantic_recall — automatic semantic long-term recall (WC-D-001).

What you'll learn:

• How to attach a semantic index to a memory.Manager with WithVectorIndex
• How memories written in one session are recalled by relevance in another
• How an agent auto-injects the top-k relevant memories each turn via WithMemoryRecall (default on when a MemoryManager has a vector index)
• How tenant scope keeps one user's memories from leaking to another

This example is fully OFFLINE: it ships a self-contained in-memory VectorStore (cosine similarity) and a deterministic hashing EmbeddingsProvider, so it runs with NO API key and NO network. In production, swap in the Qdrant adapter (storage/adapters/qdrant) and a real embeddings provider.

Run:

go run ./examples/semantic_recall/