A hospital links patient records, treatment protocols, and outcomes across 12 systems — surfacing patterns that save lives. A bank connects trading desks, compliance rules, and counterparty data — flagging risks before regulators do. A manufacturer maps suppliers, components, and production lines — detecting disruptions before they hit the floor.
An ontology defines how your data is structured — the entities, relationships, and rules of your domain. Most systems force you to define it all upfront. Scalix Prime monitors usage patterns, flags schema drift, and recommends changes — you review and approve before anything changes.
The EventCollector captures every entity creation, relationship link, property update, and query pattern — 7 event types in total. It computes a drift score that measures how well your schema matches actual data usage.
7 Event TypesBased on usage patterns, the ontology recommends schema changes — new entity types, missing relationships, property additions, index optimisations. You review and approve; the engine applies them safely with full version control and rollback.
6 Recommendation TypesA composite drift score (0.0–1.0) measures how well your schema matches actual usage. It combines unknown entity ratios, unmatched queries, and validation failure rates — flagging gaps before they become problems.
Composite Drift ScoreEvery ontology change is versioned, auditable, and reversible. Humans stay in the loop at the approval step — the engine never changes your data model without permission.
Import industry-standard ontologies out of the box. Healthcare, cybersecurity, threat intelligence, and general-purpose schemas ready to use on day one.
Three layers working together — a graph-relational engine at the foundation, a built-in ontology at the core, and enterprise applications on top.
Core Engine
Built-In Ontology
Application Layer
A native C++20 kernel where tables and graphs share the same memory space. SQL and Cypher run in the same transaction over the same data with built-in RBAC. Sub-microsecond graph traversals with Join-on-Write and Point-in-Time Recovery.
C++20 Native Kernel
Turns your codebase into a searchable knowledge graph through a 12-phase ingestion pipeline. Supports Python, TypeScript, Go, Rust, Java, and C#. Hybrid search combines BM25, vector, fuzzy, and graph signals for precise results.
Python 6 Languages
Managed PostgreSQL with vector search and physical isolation per client pod. SQL queries translate to graph operations through the shared kernel, so relational and graph workloads run over the same data without duplication.
PostgreSQL Vector SearchRelationships are computed during INSERT, not SELECT. When nodes are created, JoinRules fire automatically to create relationships — so your graph is always up to date with less than 0.5ms overhead per operation.
<0.5ms overheadVirtual Nodes map external data sources without duplication. Write-through connectors push changes back to source systems. The ReconciliationEngine handles conflicts automatically — user edits always win.
Zero DuplicationFull WAL with CRC32 integrity verification. Recover your entire database to any specific timestamp — epoch milliseconds or ISO 8601. Configurable durability modes for async or sync writes.
CRC32-Verified WALRBAC engine supports access control down to individual properties — not just tables or rows. Role hierarchy with wildcard support and 11 million permission checks per second. Row-level enforcement today, property-level filtering available via API.
11M ops/sec| Feature | Palantir Foundry | Neo4j | Scalix Prime |
|---|---|---|---|
| Architecture | Middleware over Spark | Graph-only engine | Unified graph-relational kernel |
| Query Languages | Proprietary + SQL | Cypher only | ✓ SQL + Cypher, same transaction |
| Latency | Seconds (batch-oriented) | Milliseconds | ✓ Sub-microsecond traversals |
| Privacy | Cloud-first, on-prem available | Cloud or self-hosted | ✓ Fully private, on-prem capable |
| Pricing | ✗ $1M+ / year | ✗ Enterprise licensing | ✓ Open-core, startup-friendly |
| Deployment | SaaS + on-prem enterprise | Cloud or self-hosted | ✓ Embedded, self-hosted, or cloud |
| AI Integration | AIP platform (proprietary) | Limited plugin ecosystem | ✓ NL-to-SQL, LLM routing, knowledge graphs |
| Code Intelligence | ✗ Not included | ✗ Not included | ✓ 12-phase pipeline, 6 languages |
| Built-In Ontology | ✗ Manual configuration | ✗ No ontology layer | ✓ Human-approved schema evolution + drift detection |
| RBAC | Coarse-grained | Role-based, limited | ✓ Property-level capable, 11M ops/sec |
Understand how your codebase really works. Scalix Prime maps every file, function, and dependency into a visual graph — so your team can find problems before they become outages.
Connect patient records, medications, and clinical data across departments — even when they live in separate systems. Catch dangerous drug interactions and care gaps that siloed tools miss, all while keeping data private and HIPAA-compliant.
See your entire supply chain as a connected map — from raw materials to finished products. Spot single points of failure, hidden supplier risks, and compliance gaps before they disrupt your operations.
Connect intelligence from multiple sources into a single, secure knowledge graph. Works completely offline in air-gapped environments with classification-level access control built into every query.
Map transactions, counterparties, and regulatory obligations into a single connected graph. Detect fraud patterns that span multiple accounts and institutions, and satisfy compliance requirements with built-in audit trails.
Scalix Prime is built from scratch — not bolted onto an existing database. Tables and graphs live in the same memory space, so you never have to copy data between systems. Query with SQL, Cypher, or plain English and get answers instantly.
The engine includes a built-in ontology system that lets you define the rules and structure of your domain. When data is ingested, Prime validates it against your ontology — catching inconsistencies, enforcing relationships, and ensuring quality automatically.
Every query passes through a security layer that checks access permissions at 11 million checks per second. The RBAC engine supports property-level granularity, so sensitive data stays protected without slowing anything down.
Query
Modules
Storage
Every benchmark measured on real workloads. No synthetic optimisations, no asterisks. These are the numbers the engine delivers in production.
| Operation | Throughput |
|---|---|
| 3-hop BFS (10K nodes) | 0.69 μs |
| WAL Append Record | 5.4M ops/s |
| RBAC checkAccess | 11M ops/s |
| Reconciliation (10 properties) | 511K ops/s |
Scalix RouterAn intelligent AI routing layer for enterprise clients that sits on top of Scalix Prime, turning your knowledge graph into a conversational interface.
Scalix Router connects large language models to your Scalix Prime knowledge graph — so users can ask questions in plain English and get context-aware answers grounded in your actual data. Every response is validated against the graph before it's returned.
Router handles model selection, prompt engineering, and response validation automatically. It routes each query to the right model, enriches it with graph context, and verifies the answer against your data before returning it.
Router
Prime
Security built into every layer of the stack — not bolted on as an afterthought.
One database instance per client pod. Data leaks are structurally prevented at the database layer — not just logically separated, physically separated.
TLS 1.3 in transit. WAL protected with CRC32 integrity checksums. Customer-managed encryption keys and automated key rotation on the roadmap.
Full audit trail on every operation. Designed for GDPR and HIPAA compliance. SOC 2 framework in place with certification planned as customer demand grows.
Request access and we'll set up a dedicated 2-day trial environment for your team. No credit card required.
We'll be in touch soon.