v2.0 Self-Hosted Deploy

Your Data. Your AI. Your Infrastructure.

Flora is the enterprise RAG platform that turns your internal documents into an intelligent, role-aware knowledge base — deployed entirely within your perimeter. Zero data leakage. Complete control.

Trusted by teams who take data seriously

POST /api/v1/query

How does Flora handle RBAC internally?

Status: 200 OKLatency: 124ms
{
  "text-[#8B5CF6]">"answer": "Flora endpoints enforce RBAC dynamically.",
  "text-[#8B5CF6]">"sources": ["architecture_spec_v2.pdf"],
  "text-[#8B5CF6]">"confidence_score": 0.94,
  "text-[#8B5CF6]">"language": "en"
}
Node Healthy

Enterprise Security, Built In — Not Bolted On

Data Sovereignty

Flora deploys on your infrastructure — on-premises, private cloud, or air-gapped environments. Your documents never leave your network perimeter.

GDPR & Compliance

Built-in GDPR endpoints: consent management, PII auto-detection and anonymization, right-to-erasure, and exportable audit logs. Compliance by design, not by patch.

Secrets Management

HashiCorp Vault integration for all credentials, API keys, and JWT secrets. Automatic key rotation. Zero hardcoded secrets.

SOC 2 TYPE II
GDPR COMPLIANT
ISO 27001

Everything You Need to Deploy AI Over Your Knowledge Base

Role-Based Access Control

Fine-grained, multi-tiered permissions across 8 access levels. Knowledge collections are siloed per role — public, authenticated, admin. Users only see what they're cleared to see.

Hybrid RAG Pipeline

Multi-stage retrieval combining semantic vector search, keyword matching, and BM25 with 4-signal re-ranking. Hallucination detection and multi-dimensional confidence scoring ensure verifiable answers.

Multi-Language Intelligence

Native support for English, Italian, German, and French. Automatic language detection, localized prompts, and temporal reference resolution — out of the box.

Document Lifecycle Management

Ingest PDFs, DOCX, and TXT files via API. Intelligent chunking, metadata enrichment, and batch processing with progress tracking. Full document lifecycle from upload to retirement.

Horizontal Scalability

Stateless API design with Nginx load balancing across N replicas. Redis caching, Qdrant vector clustering, and PostgreSQL replication. Kubernetes-ready with included manifests.

Production Observability

20+ Prometheus metrics, 6 pre-built Grafana dashboards, structured rotating logs. Monitor request latency, LLM token costs, cache hit rates, and vector search performance in real time.

Powered By Open-Source Local Inference

Doctor

Advanced document ingestion and parsing.

TEI (Text Embeddings Inference)

Blazing fast, local embedding generation.

Qdrant

Highly scalable, high-performance Vector Database.

vLLM

High-throughput, memory-efficient local LLM inference.

Architecture Built for the Enterprise

CLIENT LAYER
NGINX LOAD BALANCER
API REP #1
API REP #N
Qdrant
Redis
Postgres
Vault
  • Docker Compose or Kubernetes — deploy in minutes
  • 4 API replicas with Nginx load balancing included
  • Qdrant vector database for sub-second semantic search
  • Redis caching layer for intelligent response memoization
  • HashiCorp Vault for enterprise secrets management
  • Prometheus + Grafana observability stack
deploy.sh
# Deploy the full Flora stack
$ docker-compose up -d
# Or scale to meet demand
$ docker-compose up -d --scale flora-api=8

< 200ms

Average query response time

99.9%

Uptime SLA target

4 Languages

Supported out of the box

Ready to Deploy AI That Respects Your Data?

Flora runs where your data lives. No vendor lock-in, no cloud dependency, no data leakage. Start with a demo or deploy directly.

Request Demo