Typed Memory Blocks
5 structured types — FACT, PREFERENCE, EVENT, ENTITY, RELATION. Schema-validated, not untyped text blobs.
[ STRUCTURED MEMORY SDK FOR AI AGENTS ]
Memory that agents can actually reason over.
Give your AI agents persistent, structured memory they can store, query, and build context from — without external infrastructure or per-call LLM costs.
V0.4.0 — NOW AVAILABLECore Capabilities
Everything an agent needs to remember
Six foundational primitives that give your AI agents structured, searchable, and secure long-term memory.
Per-Block Encryption
AES-256 with three levels: NONE, STANDARD, SENSITIVE. Encrypt what matters, leave the rest fast.
Knowledge Graph
Built-in graph index with link, traverse, and neighbor queries. No Neo4j or external graph DB required.
Memory Decay Engine
Exponential decay with configurable rates. Prune weak memories, surface the strongest automatically.
Tamper Detection
SHA-256 hash-chained operation log. One call to verify() catches any unauthorized modification.
NEW
Async + Event Hooks
Full async API via AsyncMemBlock. Lifecycle hooks for on_add, on_update, on_delete, on_query.
Quickstart
Get started in three lines of Python.
Install from PyPI. Create a MemBlock instance. Store your first memory. No API keys, no configuration files, no setup wizard.
VIEW ON GITHUB[ PYTHON REPL ]
$ pip install memblock
>>> from memblock import MemBlock, BlockType
>>> mem = MemBlock(storage="sqlite:///memory.db")
>>> mem.store("User prefers dark mode", type=BlockType.PREFERENCE)
Block(id='blk_7f2a', type=PREFERENCE, confidence=1.0)
>>> mem.query(text_search="dark mode")
[Block(id='blk_7f2a', content='User prefers dark mode')]
>>> mem.build_context(query="user preferences", token_budget=2000)
'[PREFERENCE] User prefers dark mode (confidence: 1.00)'READY