Session Archival And Historical Indexing

via “archival memory with semantic search over documents and codebases”

Letta is the platform for building stateful agents: AI with advanced memory that can learn and self-improve over time.

Unique: Integrates archival memory as a distinct memory tier separate from working memory blocks, enabling agents to maintain both short-term context (memory blocks) and long-term knowledge (archival passages). File Processing Pipeline handles OCR, chunking, and embedding in a unified pipeline, abstracting vector database implementation details.

vs others: More integrated than standalone RAG libraries (LlamaIndex, LangChain) by tying archival memory directly to agent lifecycle and memory management; differs from simple vector search by including OCR and chunking as built-in components rather than requiring external preprocessing.