HeliosDB Edition Comparison — Nano vs Lite vs Full

HeliosDB Edition Comparison — Nano vs Lite vs Full

Three editions, one wire-protocol family. This doc clarifies the positioning of each edition, the architectural differences, and which features ship where.

Versioning note: each edition has its own version line. Nano is at v4.41.0, Lite is at v3.6.0, Full is at v8.3.0. The version numbers are independent — a higher number does not imply a later release date.

---

At a Glance

Dimension	HeliosDB Nano (v4.41.0)	HeliosDB Lite (v3.6.0)	HeliosDB Full (v8.3.0)
Architecture	Single embedded engine	Single embedded engine	Distributed cluster
Primary target	Embedded / single-process / AI agents	Embedded + self-hosted server, AI on edge	Frontier distributed enterprise database
Deployment	Library + standalone server (~32 MB binary)	Library + standalone server	Server cluster (Docker / K8s / Helm / Terraform / Ansible)
License	Apache-2.0	SSPL-1.0	SSPL-1.0
Storage engine	RocksDB	RocksDB + native storage engine	Self-developed LSM storage engine
Distribution model	Tier 1–3 HA (warm standby / multi-primary / sharding)	Tier 1–3 HA + sharding	Full Raft consensus + multi-region active-active + sharding
Wire protocols	PostgreSQL + MySQL (+ REST API + Realtime WS)	PostgreSQL + Oracle	14 protocols (PG / MySQL / MongoDB / Redis / Cassandra / ClickHouse / Oracle / SQL Server / DB2 / Snowflake / Databricks / Pinecone / HTTP / GraphQL)
Backend-as-a-Service	✅ Built-in (Auth, REST, Realtime, RLS, Swagger UI)	❌	❌ (use the REST / GraphQL APIs)
HeliosProxy (standalone, Apache-2.0)	✅ Works with Nano (PostgreSQL wire)	✅ Works with Lite	✅ Works with Full (optional — the cluster also has its own built-in routing)

---

Feature Matrix

Core Database

Feature	Nano	Lite	Full
PostgreSQL wire protocol	✅	✅	✅
MySQL wire protocol	✅	❌	✅
MongoDB wire protocol	❌	❌	✅
Redis protocol	❌	❌	✅ (134+ commands incl. AI-native)
Oracle wire protocol	❌	✅	✅
SQL Server / DB2 / Cassandra / ClickHouse / Snowflake / Databricks / Pinecone	❌	❌	✅
HTTP / REST + GraphQL	✅ (BaaS)	❌	✅
Embedded mode (in-process)	✅	✅	⚠️ Possible via library, not the typical deployment
Single-binary deployment	✅	✅	✅
Standalone server mode	✅	✅	✅ (primary deployment mode)
WordPress drop-in (stock wpdb)	✅	❌	✅

Indexing & Search

Feature	Nano	Lite	Full
HNSW vector search	✅	✅	✅
Product Quantization (8–16x compression)	✅	✅	✅
BM25 full-text search	✅	✅	✅
Hybrid (BM25 + vector) via SQL	✅	✅	✅ (via hybrid_search SQL function in v7.4+)
GIN-style FTS DDL	✅	✅	✅
Geospatial (PostGIS-compatible)	❌	❌	✅

Transactions & Concurrency

Feature	Nano	Lite	Full
ACID transactions	✅	✅	✅
MVCC	✅	✅	✅
Savepoints	✅	✅	✅
Distributed transactions	❌	⚠️ Limited	✅
Distributed deadlock detection	❌	❌	✅

Time-Travel & Branching

Feature	Nano	Lite	Full
Git-like branches (CREATE / USE / MERGE / DROP BRANCH)	✅	✅	✅
Copy-on-write branch creation	✅ (microsecond)	✅ (microsecond)	✅ (~555 µs)
AS OF TIMESTAMP/TRANSACTION/SCN	✅	✅	✅ (SQL:2011 temporal)
Selective branch sync to remotes	✅ (v3.19)	❌	⚠️ via cluster replication

Distributed & HA

Feature	Nano	Lite	Full
Tier-1 warm standby (default-on)	✅	✅	✅
Tier-2 multi-primary (vector clocks)	✅	✅	✅
Tier-3 sharding	✅ (consistent hash ring)	✅	✅ (hash / range / geohash + rebalancer)
Raft consensus	❌	❌	✅ (metadata + cluster)
Multi-master CRDT (<50ms global writes)	❌	❌	✅
Cross-region active-active	❌	❌	✅
Multi-region	❌	❌	✅
Adaptive ML query routing	❌	❌	✅
Multi-armed bandit load balancer	❌	❌	✅
Backup & disaster recovery	⚠️ Dump/restore	⚠️ Dump/restore	✅ Comprehensive
Point-in-time recovery (PITR)	❌	❌	✅

AI / ML

Feature	Nano	Lite	Full
NL queries (basic)	✅	✅	✅
RAG primitives (chunking, retrieval)	✅	✅	✅
Code-graph (LSP via SQL, tree-sitter)	✅ (opt-in)	❌	❌
Graph-RAG (universal node/edge schema)	✅ (opt-in)	❌	❌
Cognitive agents (5: optimizer / index advisor / troubleshooter / tuner / schema mgr)	❌	❌	✅
In-database ML training (SQL-based)	❌	❌	✅
Anomaly detection	❌	❌	✅
Time-series forecasting (ARIMA / Prophet / LSTM)	❌	❌	✅
AutoML hyperparameter tuning	❌	❌	✅
ML model registry	❌	❌	✅
Neural query planner	❌	❌	✅
ML-based intelligent data tiering	❌	❌	✅
Privacy-preserving ML (differential privacy)	❌	❌	✅
Federated learning (100+ nodes)	❌	❌	✅
Conversational BI (multi-turn)	❌	❌	✅
NL2Graph (GraphRAG)	❌	❌	✅
AI schema architect	❌	❌	✅
Embeddings (OpenAI / Cohere / Voyage / local)	✅ (local + HTTP)	✅	✅

Multi-Model

Feature	Nano	Lite	Full
Relational	✅	✅	✅
Document (MongoDB-compatible)	⚠️ JSONB only	⚠️ JSONB only	✅ Full document store
Graph engine (in-memory)	✅ (basic)	✅ (basic)	✅ (with optimization)
Cypher / GQL (ISO/IEC 39075)	❌	❌	✅
Geospatial (PostGIS)	❌	❌	✅
Probabilistic queries	❌	❌	✅
Time-series	⚠️ via SQL	⚠️ via SQL	✅ Native

Lakehouse & Federation

Feature	Nano	Lite	Full
Apache Iceberg	❌	❌	✅
Delta Lake	❌	❌	✅
Unified catalog	❌	❌	✅
Federated query (AWS / Azure / GCP)	❌	❌	✅
Data sovereignty	❌	❌	✅

Backend-as-a-Service (BaaS)

Feature	Nano	Lite	Full
REST API (PostgREST-compat, 19 filter operators)	✅	❌	✅ (general REST + admin)
Auth Service (signup / token / refresh)	✅	❌	⚠️ External auth (SSO / LDAP)
OAuth2 (Google / GitHub) with PKCE	✅	❌	❌
Realtime WebSocket (Phoenix protocol)	✅	❌	❌
Swagger UI (auto-generated)	✅	❌	✅ (admin endpoints)
Row-Level Security (RLS) on REST	✅	✅	✅

Encryption / Security / Compliance

Feature	Nano	Lite	Full
AES-256-GCM TDE	✅	✅	✅
FIPS 140-3 (AWS-LC)	✅ (opt-in)	❌	⚠️ Enterprise security build
Zero-Knowledge Encryption (ZKE)	❌	✅	❌
SCRAM-SHA-256	✅	✅	✅
TLS / SSL (PG + MySQL)	✅	✅	✅
RBAC	⚠️ Basic	⚠️ Basic	✅ v2
ABAC	❌	❌	✅
SSO (SAML / OIDC)	❌	❌	✅
LDAP / Active Directory	❌	❌	✅
Post-quantum cryptography (CRYSTALS-Kyber)	❌	❌	✅
Homomorphic encryption (CKKS)	❌	❌	✅
Multi-cloud secrets management	❌	❌	✅
Compliance framework (SOC2 / HIPAA / GDPR / PCI-DSS / FIPS)	⚠️ Manual	⚠️ Manual	✅ Built-in
Audit logging	✅ Tamper-proof SHA-256 chain	✅ Built-in	✅ + Blockchain audit trail + Lineage

Multi-Tenancy

Feature	Nano	Lite	Full
Row-Level Security (policy-based)	✅	✅	✅
Tenant quotas	⚠️ Basic	✅	✅ Built-in
Enterprise multi-tenancy (1000+ tenant isolation)	❌	❌	✅
Connection pooling (session / transaction / statement)	⚠️ Basic (full pooling via HeliosProxy)	✅ via HeliosProxy	✅
Customer onboarding (signup / trials / portal)	❌	❌	✅

Caching & Performance

Feature	Nano	Lite	Full
Result cache (LRU)	✅	✅	✅
Schema cache	✅	✅	✅
Adaptive Radix Tree (ART) PK index	✅	✅	✅
Bloom filters	✅	✅	✅
RL-based caching	❌	❌	✅
4-tier edge cache (browser → DB)	❌	❌	✅
Global distributed cache (cross-region)	❌	❌	✅
Intelligent prefetching	❌	❌	✅
Quantum-inspired optimizer	❌	❌	✅ (fallback-only by default)
GPU acceleration (CUDA / ROCm)	❌	❌	✅

Edge & Streaming

Feature	Nano	Lite	Full
Edge computing (compute / sync / AI / embedded-cloud)	❌	⚠️ Edge/geo caching via HeliosProxy	✅
Streaming ETL (with AI data mapping)	❌	❌	✅
Webhooks (HTTP edge functions)	❌	❌	✅
Cron scheduler (edge function scheduling)	❌	❌	✅
CDC (change data capture)	⚠️ Core (persistence WIP)	⚠️ Core	✅ Full pipeline
Tenant replication	❌	❌	✅

WASM & Extensibility

Feature	Nano	Lite	Full
WASM stored procedures (Rust / JS / Python / Go)	❌	❌	✅
WASM SDK + runtime + macros	❌	❌	✅
Database triggers	✅ Built-in	✅ Built-in	✅ + WASM-based
Python WASM bindings	❌	❌	✅

Observability

Feature	Nano	Lite	Full
tracing structured logs	✅	✅	✅
EXPLAIN ANALYZE	✅	✅	✅ (7 formats: Text / JSON / YAML / XML / HTML / GraphViz / Mermaid)
OpenTelemetry distributed tracing	❌	❌	✅
Prometheus metrics export	⚠️ Basic	⚠️ Basic	✅
Visual query debugger	❌	❌	✅
Business metrics & BI dashboards	❌	❌	✅

MCP (AI Agents)

Feature	Nano	Lite	Full
Native MCP JSON-RPC server	✅ (opt-in mcp-endpoint feature, v3.18+)	❌	✅ (v7.4+)
Tool count	16 (10 DB-backed + 6 RAG)	—	14 (10 base + 4 Full-only)
Full-only tools	—	—	heliosdb_cypher_query, heliosdb_nl2sql, heliosdb_lakehouse_query, heliosdb_index_advisor
Transports	WebSocket / stdio / SSE	—	stdio

Research & Innovation (Full-only)

Feature	Nano	Lite	Full
Quantum-inspired query optimization	❌	❌	✅
GPU acceleration (CUDA / ROCm)	❌	❌	✅
Genomics processing	❌	❌	✅
Neuromorphic computing (spiking NN)	❌	❌	✅
Blockchain-CRDT hybrid sync	❌	❌	✅
Holographic visualization (WebXR)	❌	❌	✅

---

Choosing an Edition

Choose Nano when…

• You’re embedding a database into a single-process application (desktop, mobile, IoT, edge).
• You want PostgreSQL + MySQL + REST + Realtime + Auth + RLS in one ~32 MB binary.
• You’re building an AI-agent backend and want a native MCP endpoint.
• You’re hosting WordPress and want a drop-in MySQL replacement (zero plugin / db.php changes).
• You need code-graph (LSP via SQL) or Graph-RAG for AI coding tools.

Choose Lite when…

• You want PostgreSQL-compatible embedded + server with Zero-Knowledge Encryption (ZKE).
• You’re deploying to edge nodes that need local DB + intelligent routing + AI-aware connection multiplexing.
• You want the lightest enterprise feature set (RLS, audit, multi-tenancy v2) without the full distributed stack.

Choose Full when…

• You need Raft consensus + multi-region active-active + sharding.
• You’re consolidating multiple databases (PG / MySQL / MongoDB / Redis / Cassandra / Oracle) into one platform.
• You need lakehouse integration (Iceberg + Delta) alongside transactional SQL.
• You want federated learning across organizations (100+ nodes, privacy-preserving).
• You need conversational BI, cognitive agents, AutoML, anomaly detection, forecasting built-in.
• You’re under SOC2 / HIPAA / GDPR / PCI-DSS / FIPS — Full ships the compliance framework.
• You need post-quantum cryptography or homomorphic encryption for regulated data.
• You want quantum-inspired query optimization for complex OLAP plans (fallback-safe).
• You need a plugin ecosystem with REST / GraphQL admin endpoints, 14 native protocol adapters, and SSO / LDAP integration.

---

HeliosProxy — Works With Every Edition

HeliosProxy is a standalone, Apache-2.0 product — it is not tied to any one edition. It is a PostgreSQL-wire proxy that runs in front of HeliosDB Nano, Lite or Full, or in front of stock PostgreSQL, and provides:

• Connection pooling with session / transaction / statement modes
• Read/write splitting, health checking, circuit breaking and failover
• A transaction journal with operator-driven replay
• Optional capabilities behind feature flags: WASM plugins, query caching, query analytics, multi-tenancy and more

Source and releases: https://github.com/HeliosDatabase/HeliosDB-Proxy · Documentation: HeliosProxy docs

With Full, the cluster also includes its own built-in routing component (transparent write routing, transaction replay and failover orchestration as part of the cluster), so a Full deployment does not require HeliosProxy. You can still place HeliosProxy in front of a Full cluster when you want its pooling, policy or observability features.

---