SPT Examples

Practical examples demonstrating SPT capabilities for vector intelligence, agent coordination, and distributed systems.

SptDB (Vector Database)

Work with SPT's vector database and memory operations.

Example	Description
Batch Query	Execute multiple vector queries with throughput metrics
Batch Store	Store multiple patterns/vectors in bulk with progress tracking
Cached Query	Query cache layer with TTL for repeated queries (2000x speedup)
Connection Pool	Connection pool manager for concurrent database access
Perf Monitor	Performance monitor that tracks query/store metrics

Swarm Coordination

Multi-agent swarm orchestration and QUIC transport.

Example	Description
Multi-Agent Deployment	Full swarm deployment with hierarchical topology and 4 agent types
QUIC Server Coordinator	QUIC server as central coordinator for distributed agent swarms
QUIC Swarm Coordination	Distributed code review and refactoring pipeline over QUIC

ReasoningBank

Self-learning reasoning system with SAFLA capabilities.

Example	Description
Benchmark	Performance benchmark suite for storage, query, and optimisation
Learning Demo	Self-learning with semantic pattern storage and cross-domain queries
Optimize	Optimisation suite: VACUUM, caching, batch ops, performance tips

NATS Integration (Legacy)

Legacy NATS examples. The core transport has been replaced by SPT Wire (spt-wire crate, QUIC-based). NATS remains available as a SmartConnector for bridging to external systems. See ADR-088.

Example	Description
Swarm Pub/Sub	Pub/sub patterns for agent status, task distribution, memory coordination
JetStream Task Queue	Persistent task queues with at-least-once delivery and consumer groups
KV Agent State	Key-value store for real-time agent state management with watches

Prerequisites

• Node.js 18+
• For swarm examples: ANTHROPIC_API_KEY environment variable
• For legacy NATS examples: nats-server running locally
• For ReasoningBank: npx spt CLI available