Deploy intelligent AI assistants powered by your knowledge base. Any LLM provider, multi-agent orchestration, enterprise-grade security — purpose-built for enterprise.
Multi-agent AI system
A team of specialised agents — support, sales, analytics, scheduling, CRM, plus industry specialists — works in parallel on every query. Routing is automatic, answers are synthesised into one coherent response.
Any LLM provider
OpenAI, Anthropic, Azure, Groq, Ollama, vLLM, and any OpenAI-compatible endpoint — or bring your own model. Automatic fallback across the provider chain so a single outage doesn't take the widget down.
Knowledge base
Upload PDF, DOCX, CSV, TXT, or Markdown. Or scrape your website automatically. Vector search finds the right answers in milliseconds. AI learns in minutes.
Embeddable widget
Fully customisable chat widget with your brand colours, logo, and messaging. One script tag. Works on any website, CMS, or web app — no rebuild required.
Enterprise security
Tenant isolation via PostgreSQL Row-Level Security, real-time PII detection and masking, prompt-injection firewall. GDPR-compliant, HIPAA-ready; logs structured for SOC 2 and ISO 27001 audits.
Cost optimisation
Semantic routing keeps simple queries on cheap local models and reserves frontier LLMs for complex work. Tiered routing is the cost-control lever; actual savings depend on your traffic shape.
Tech stack
- FastAPI
- LangGraph
- PostgreSQL
- Qdrant
- Redis
- OpenTelemetry
- React
- Docker
~300ms
SLM time-to-first-token
6
Widget languages
HA
Multi-provider failover
RLS
Per-tenant isolation
Deployment
Live in under 10 minutes
From zero to a fully deployed, knowledge-powered AI assistant in three steps.
- 01
Upload your knowledge
Drop in your PDFs, Word docs, spreadsheets, or paste URLs. Our pipeline automatically chunks, embeds, and indexes everything for lightning-fast retrieval.
- 02
Configure your widget
Match your brand in minutes. Set colours, name your assistant, write a custom greeting, choose your LLM provider, and define which agents handle which queries.
- 03
Deploy with one line
Paste a single <script> tag into your site. The widget loads asynchronously — zero impact on your page speed.
Security & Compliance
Built for enterprise from day one
Security is not a feature — it is the foundation every component is built on.
Tenant Isolation
PostgreSQL Row-Level Security ensures absolute data separation between tenants. No tenant can ever see another's data — enforced at the database engine level.
PII Auto-Detection & Masking
Presidio-powered real-time detection of SSNs, credit cards, emails, phone numbers, and 30+ other PII types. Automatically masked before any LLM sees your data.
Prompt Injection Firewall
Multi-layer defence against prompt injection, jailbreaking, and adversarial inputs. Semantic analysis catches attacks that keyword filters miss.
Tamper-Evident Audit Logs
Every action is logged with HMAC-SHA256 signatures in a hash-chained audit trail. Evidence artefacts map to SOC 2, ISO 27001, and HIPAA control requirements.
Widget Variants
One platform, every surface
Deploy the right widget for every context — from public websites to internal portals to native mobile apps.
- Compact
Floating Bubble
380×600 px. The classic bottom-right chat bubble. Zero friction for public websites and e-commerce.
- Tablet
Assistant Panel
768×1024 px with sidebar. Designed for internal portals and knowledge management.
- Mobile
Fullscreen App
100vw × 100vh. Native-feel fullscreen overlay for mobile apps and PWAs.
Watch
See the platform in action
A visual walkthrough of key capabilities, from knowledge ingestion to intelligent multi-agent responses.
Listen
Hear the platform explained
A deep dive into the architecture, security, and intelligence that powers EinCoreRAG.
Inside EinCoreRAG
Platform Deep Dive
Industry playbooks
Pre-built for verticals that need it most
Each playbook comes with industry-specific agents, workflows, compliance posture, and example conversations — backed by a live pilot tenant you can chat with right now.
- 🏥
Clinic in a Box
AI front desk for medical practices. Books appointments, triages symptoms, answers insurance questions. HIPAA-ready.
- 🛍️
E-commerce Concierge
Shop assistant that answers product questions, processes refunds with manager approval, recovers carts. Stripe integration.
- 🛎️
Hotel Front Desk
Multilingual concierge for hotels. Reservations, amenities, local guide, VIP routing. GDPR-ready for EU guests.
Enterprise
Built exclusively for enterprise
Tailored deployment, dedicated support, and complete control. Let us build the right solution for your organisation.
Talk to salesFAQ
Frequently asked questions
Will it make up answers?
No. Responses are retrieved from your own indexed knowledge base, and routing and synthesis stay grounded in those sources rather than the model's memory.
Is each customer's data actually isolated?
Yes. Tenant separation is enforced in PostgreSQL Row-Level Security at the database engine level, so no tenant can ever see another's data.
We can't send our data to a US cloud — what are our options?
Deploy the EU-sovereign path on Mistral and Qdrant with EU data residency, or self-host the inference layer. PII is auto-masked before any model sees it, and tamper-evident audit logs map to SOC 2 / ISO 27001 evidence.
Have a project in mind?
Tell us what you want to build. We respond within one business day.