SPINE Demos Memory Explorer

Memory Explorer

Interactive visualization of SPINE's 7-tier memory architecture with unified facade and verdict routing.

The 7-Tier Memory Stack

T7
GraphMemory
Relationship traversal + analytics
T6
DeepMemoryStore
PostgreSQL + pgvector long-term storage
T5
EpisodicMemory
Goal-based episode recording + recall
T4
VectorStore
Semantic + keyword hybrid search
T3
EphemeralMemory
Session context with time-based decay
T2
Scratchpad
Working memory / task notes
T1
KVStore
Fast key-value pairs by namespace
Click a tier in the stack to explore it

MemoryFacade: Unified Query Flow

MemoryFacade.search() fans out a single query across all 7 tiers, normalizes scores, and returns ranked results.

MemoryFacade
Score normalization + tier weighting
T1
--
T2
--
T3
--
T4
--
T5
--
T6
--
T7
--

VerdictRouter: Decision Routing

After task evaluation, the VerdictRouter directs results to the appropriate memory tier based on the verdict.

AgenticLoop Evaluator
Click a verdict above
VerdictRouter
route(verdict, task, result)
ACCEPT
Store as Episode
T5: EpisodicMemory
REVISE
Update revision notes
T2: Scratchpad
REJECT
Log rejection reason
T3: EphemeralMemory

Tier Comparison

Tier Name Backend Scope Persistence
1 KVStore SQLite / File Namespace-scoped Durable
2 Scratchpad In-memory list Task-scoped Volatile
3 EphemeralMemory In-memory + decay Session-scoped TTL Decay
4 VectorStore LanceDB / Keyword Collection-scoped Durable
5 EpisodicMemory SQLite + FTS5 Goal-scoped Durable
6 DeepMemoryStore PostgreSQL + pgvector Project-scoped Durable
7 GraphMemory PostgreSQL + NetworkX Cross-entity Durable

Cross-Cutting Components

F

MemoryFacade

Unified search across all 7 tiers. Score normalization and tier weighting merge results from different backends into a single ranked list.

facade.search("query", top_k=10)
M

FederatedMemory

Read-only cross-project memory federation. Parallel fan-out to remote Minna MCP servers with budget controls and provenance metadata.

fed.search("query", limit=10)
V

VerdictRouter

Routes evaluator verdicts to the appropriate tier. Accept stores episodes, Revise updates scratchpad, Reject logs to ephemeral memory.

router.route(verdict, task, result)

How the Memory System Works

SPINE's memory architecture uses 7 tiers to match different temporal scopes and access patterns. Fast lookups use Tier 1 (KVStore), session-scoped observations use Tier 3 (Ephemeral), and long-term semantic knowledge lives in Tier 6 (DeepMemoryStore) with PostgreSQL + pgvector.

facade = MemoryFacade(kv, scratchpad, ephemeral, vector, episodic, deep_store, graph_memory)
results = facade.search("authentication", top_k=10)
# Returns ranked results from whichever tiers have relevant data

The MemoryFacade provides a single search interface across all tiers. The VerdictRouter closes the feedback loop by directing task outcomes to the appropriate tier based on the evaluator's verdict. FederatedMemory extends reach across project boundaries via parallel MCP server queries.