We use cookies to improve your experience. By using this website you agree to our Cookie Policy

Security Compliance
Smabbler Galaxia Security Overview
1. Data Processing Security
  • All data processing in Galaxia occurs entirely in-memory within Smabbler’s proprietary infrastructure (AWS). No external AI models, vector databases, or third-party embeddings are used.
  • Processing Workflow (Fully Contained & Deterministic)
Stage
Description
Security Note
1. NLP Analysis
Input text is processed using Galaxia’s proprietary natural language engine.
No external APIs or LLMs are called.
2. Semantic Enrichment
Data is contextualized using Galaxia’s built-in multilingual hypergraph knowledge base.
Uses only internal ontologies and taxonomies.
3. Graph Creation
A custom knowledge graph is generated from the user’s data.
Resulting graph is isolated and owned entirely by the user.
4. In-Memory Execution
The graph is loaded and executed in an in-memory runtime.
No data leaves Smabbler infrastructure.
  • No neural networks
  • No external embeddings
  • No third-party processing
  • Fully deterministic and explainable pipeline
2. Data Storage & Deletion Policy
  • Source files are automatically deleted within 24 hours of upload (user may delete data earlier at any time).
  • The resulting knowledge graph is retained only in the user’s account and can be permanently deleted on demand.
  • No data is ever shared across accounts or used for model training.
  • All data is stored in encrypted form (at rest and in transit).
3. Chat Interface & Generative Layer
  • Galaxia includes an optional chat interface powered by AWS Bedrock LLMs (e.g. Claude Sonnet, Haiku).
  • These models are used only for natural language interaction, not for data processing.
  • In line with AWS Bedrock policy, no user data is used for training or retained by model providers.
  • Users can disable the chat interface entirely and use only the hypergraph query engine if required for compliance.
4. Access Control & Account Security
  • Each knowledge graph is fully isolated and accessible only to the account owner.
  • Access is secured with industry-grade encryption, role-based permissions, and via dedicated user keys.
5. Compliance-Ready Architecture
  • Galaxia’s explainable hypergraph model enables auditability and GDPR compliance by design:
  • Every node and relationship has full provenance tracking.
  • Users can delete any data at any time to support the right to be forgotten.
  • No opaque embeddings mean data is always human-readable and traceable.
  • Summary: Why Enterprises Trust Galaxia
Feature
Galaxia
Data Storage
In-memory, CPU-native
Data Use for Training
Never
Explainability
Full provenance
Data Residency
EU / Private
Deletion & Control
User-driven
Feature
Traditional LLM Systems
Data Storage
Third-party GPU clouds
Data Use for Training
Often unclear
Explainability
Black box
Data Residency
Restricted
Deletion & Control
Not guaranteed
Feature
Galaxia
Traditional LLM Systems
Data Storage
In-memory, CPU-native
Third-party GPU clouds
Data Use for Training
Never
Often unclear
Explainability
Full provenance
Black box
Data Residency
EU / Private
Restricted
Deletion & Control
User-driven
Not guaranteed
Galaxia PlatformBeta AccessPublic Graph ModelsInside Galaxia
InsightsSolutionTechnologyResearch InsightsApplicationsPerspectivesCase Studies
For DevelopersDocumentationAPI AccessLlama IndexLangFlow
Contact Us
Term & PrivacyNCBiR Research Grant

Copyright © 2025 Smabbler ®. All rights reserved