Aurix Datastore

Your Data, Semantically Indexed,
Never Leaving Your Network

A cognitive data layer that understands your content at ingest. LLM-assisted semantic indexing makes every document, database, and API queryable by meaning, not just keywords.

Enterprise Pricing Contact Sales
100%
On-Premise
AES-256
Encryption at Rest
RAG
Retrieval Pipeline
RBAC
Access Control
How It Works

LLM-assisted indexing at ingest

Traditional search indexes keywords. Aurix uses an LLM to understand the semantic meaning of your content at the moment it's ingested, creating a rich knowledge graph that enables contextual retrieval.

1

Ingest

Documents, database records, and API responses are fed into the Aurix pipeline. Connectors handle format conversion automatically.

2

Understand

An LLM analyzes each chunk, generating semantic embeddings, entity extractions, relationship mappings, and contextual summaries.

3

Retrieve

Queries are matched against semantic meaning, not just keywords. The RAG pipeline returns the most relevant context for your agents.

RAG Pipeline

Context-aware retrieval for agent workflows

Aurix powers the retrieval-augmented generation (RAG) pipeline for Aura Workshop agents. When an agent needs context, Aurix delivers the most relevant documents, code snippets, and data based on semantic similarity.

  • Semantic Search — Find documents by meaning, not just keyword matching
  • Hybrid Retrieval — Combines vector similarity with BM25 keyword scoring
  • Contextual Chunking — Smart document splitting that preserves semantic boundaries
  • Re-Ranking — LLM-based re-ranking to surface the most relevant results
  • Citation Tracking — Every retrieved chunk links back to its source document and location
  • Agent Integration — Seamless integration with Aura Workshop agent workflows
Retrieval Pipeline
Query Understanding semantic analysis
Vector Search cosine similarity
Keyword Search (BM25) term frequency
Re-Ranking LLM-scored
Context Delivery to agent
Data Sources

Connect everything. Index once. Query forever.

Aurix ingests data from multiple sources through its connector framework. Each connector handles format-specific parsing and normalization.

Documents

PDF, DOCX, XLSX, PPTX, Markdown, plain text, HTML. Automatic format detection and content extraction.

Source Code

Git repositories, individual files, or entire codebases. Language-aware parsing preserves function and class boundaries.

Databases

PostgreSQL, MySQL, SQLite, MongoDB. Schema-aware indexing captures table relationships and data semantics.

REST APIs

Connect to internal APIs and index response data. Scheduled polling keeps the index fresh as data changes.

Email & Messages

IMAP email, Slack exports, and communication archives. Thread-aware indexing preserves conversation context.

File Systems

Watch directories for changes. New and modified files are automatically re-indexed in near real-time.

Security

Enterprise-grade security. Your infrastructure. Your control.

Aurix is built for organizations that handle sensitive data. Every security decision defaults to the most restrictive option. Your data never leaves your infrastructure.

  • Encryption at Rest — AES-256-GCM encryption for all stored data and indexes
  • Encryption in Transit — TLS 1.3 for all internal communication between components
  • Access Controls — Per-collection, per-document, and per-field access restrictions
  • Audit Logging — Every query, every access, every modification logged with timestamps
  • Network Isolation — Runs entirely on your network. No external calls. No phone home.
  • Data Retention Policies — Configure auto-deletion schedules per collection or data source
Security Stack
AES-256-GCM Encryption at Rest
TLS 1.3 Encryption in Transit
Role-Based Access Control (RBAC)
Per-Document Access Policies
Complete Audit Trail
Zero external dependencies.
Enterprise Features

Built for teams. Built for scale.

Aurix Datastore includes enterprise features that support multi-team deployments, compliance requirements, and large-scale data operations.

Multi-Tenant

Isolated data stores per tenant with independent indexes, access controls, and retention policies.

RBAC

Role-based access control with predefined roles (Admin, Editor, Viewer) and custom role creation.

Horizontal Scaling

Distribute data and query load across multiple nodes. Add capacity without downtime.

Incremental Indexing

Only re-index changed content. Efficient delta updates minimize compute cost for large datasets.

Monitoring Dashboard

Real-time metrics for index size, query latency, ingestion throughput, and storage utilization.

Dedicated Support

Enterprise customers get dedicated engineering support, custom connector development, and SLA guarantees.

Ready to own your data infrastructure?

Aurix Datastore is available as part of the Enterprise plan. Contact us for a demo and custom pricing.

View Enterprise Pricing Contact Sales