Book a Demo
LOCAL AI · ZERO DATA LEAKAGE · CONTROLLED EARLY ACCESS

Your regulations.
Your server.
Your answers.

EngiIntel deploys on your own infrastructure. Engineers query regulations, manage a built-in knowledge base, track assets, generate incident reports, automate workflows with n8n and run autonomous AI agents — with zero data leaving your server.

GDPR compliant by architecture
Zero external API calls
Works fully offline
13× Anthropic AI Certified
ENGIINTEL AGENT · ACTIVE SESSION
AGENT: "File incident report for LIFT-CM14-03 fault"

⟳ [1/4] search_documents → EN 81-20 §9.7 · 3 results · 1.8s
⟳ [2/4] get_asset → LIFT-CM14-03 · operational
⟳ [3/4] create_task → EU-47 · priority: high
⟳ [4/4] generate_incident_report → IR-2026-047

✓ AGENT COMPLETE · 4 steps · 3.4s · fully local

"Incident IR-2026-047 created and linked to
LIFT-CM14-03. Deadline: 4h per EN 81-20 §9.7.
Task EU-47 assigned. Wiki page published."

TOOLS: search · asset · task · report · local llm
0ms
Cloud Latency
100%
Data Stays On-Premise
5+
Document Connectors
Queries — No API Costs
EARLY ACCESS Early access is open — request a slot and we'll onboard you personally.
Request Beta Access →
Engineers spend 30 minutes searching for an answer that should take 30 seconds.

Regulated industries run on documentation. But finding the right clause in the right regulation, across hundreds of PDFs, costs your team hours every day — hours that could be spent on actual engineering work.

73%
of engineering time spent searching for information, not creating it
4h
lost per engineer per week to manual document search
€960
recovered per engineer/month with EngiIntel at €80/h
0
documents leave your server. Ever.
Capabilities

Everything your engineers need,
nothing your legal team fears.

Built for regulated industries where document accuracy, traceability, and data confidentiality are non-negotiable.

🔍
Natural Language Queries
Ask questions in plain Spanish or English. Get precise answers with exact page citations from your technical documentation.
RAG · HYBRID SEARCH
🔒
100% On-Premise AI
Runs entirely on your hardware. No data sent to OpenAI, Google or any third party. Full GDPR compliance by design.
LOCAL LLM · OLLAMA
📄
OCR for Scanned Documents
Processes scanned PDFs and image-based documents. Legacy manuals and old regulations work just as well as native PDFs.
TESSERACT · VISION AI
🔗
Document Connectors
Automatic sync from SharePoint, Google Drive and Confluence. Documents stay up to date without manual uploads.
SHAREPOINT · GDRIVE · CONFLUENCE
👥
Department Access Control
Each department sees only its own documents. Admins manage users, roles and permissions from a dedicated panel.
RBAC · MULTI-TENANT
📊
Audit Logging & Compliance
Every query, upload and action is logged with user, timestamp and response time. Full traceability for audits.
ISO READY · IMMUTABLE LOGS
🏗️
Hierarchical Knowledge Base
Organise documents in folders with AI context instructions. Query by department, project or regulation type — scope narrows automatically.
SMART SCOPING
📈
ROI & Usage Analytics
Track hours saved and recovered engineering time per team. One-click executive PDF report for quarterly reviews.
ROI DASHBOARD · PDF EXPORT
🔐
SSO & Enterprise Auth
Microsoft Entra ID SSO out of the box. GitHub OAuth fallback. RBAC scoped per department.
ENTRA ID · OAUTH · RBAC
🛡️
PII Detection & GDPR
Automatic PII scanning across all indexed content. GDPR right-to-erasure endpoint included.
PII SCAN · GDPR ERASURE
🔌
Public API & Webhooks
Versioned REST API with per-tenant API keys. Webhooks for document indexing, PII and confidence alerts.
REST API · WEBHOOKS · HMAC
One-Command Deployment
Fully containerised with Docker. IT runs one command and the entire stack — AI, database, app — is live in minutes.
DOCKER · SELF-HOSTED
📋
Excel & Multi-Format Support
Ingest Excel workbooks alongside PDFs. Technical specifications, test results and data sheets all searchable.
XLSX · MULTI-FORMAT
🔄
Document Comparison
Semantically compare two documents side by side. Identify differences and conflicts between regulation versions instantly.
SEMANTIC DIFF · STREAMING
🧠
Query Result Cache
Near-identical questions return instantly from an in-memory LRU cache — no inference needed. Cosine similarity (≥98%) detects duplicates per organisation, with a 1-hour TTL.
LRU CACHE · COSINE SIM
🔂
Document Deduplication
SHA-256 hash check at upload time prevents the same file being indexed twice. Zero silent vector store pollution — exact match detected with a clear error showing the original document.
SHA-256 · DEDUP
📐
Latency Percentiles p50/p95/p99
The admin dashboard shows full response time breakdowns — p50, p95 and p99 — not just the average. Track tail latency accurately before making any SLA commitments.
P95 · LATENCY · SLA
📖
Wiki & Knowledge Base
Built-in wiki with Markdown editor, version history, templates and page types — procedures, checklists, meeting notes and incident reports. All pages are RAG-indexed automatically and searchable via AI queries.
WIKI · VERSION HISTORY · RAG
🏗️
Asset Registry & Incident Reports
Register and track physical assets — elevators, escalators, equipment. Link incidents to assets, track regulatory response deadlines and auto-generate AI-written incident reports directly from task data.
ASSETS · INCIDENTS · AUTO-REPORT
🧩
Cross-Module AI Context
Pin any record — wiki page, asset, incident, task — to the persistent context panel. Ask the AI one question that synthesises all pinned items simultaneously. The AI also passively tracks what you're viewing across every module.
CONTEXT PANEL · CROSS-MODULE
🤖
AI Agent Mode
Switch to Agent Mode and let the AI autonomously execute multi-step workflows: search documents, create tasks, query assets, generate incident reports — a full ReAct loop with live step-by-step reasoning visible in the UI.
REACT LOOP · AUTONOMOUS · SSE
🔄
n8n Workflow Automation
Create, manage and execute n8n workflows directly from AI Agent Mode. Automate document processing, approval routing, notifications and integrations with your existing tools — all without leaving EngiIntel.
N8N · MCP · AUTOMATION
Modular Platform

Build your stack.
Pay for what you use.

Every module works standalone or combined. Start with the RAG core and add capabilities as your team grows — no bloated suites, no forced bundles.

🧠
Core — Document Intelligence Engine
The mandatory base of every EngiIntel deployment. RAG pipeline with hybrid search, zero-hallucination answers with page-level citations, document connectors (SharePoint, Google Drive, Confluence), hierarchical folder system, role-based access control, OCR for scanned docs, and full audit logging. Your documents never leave your server.
RAG · HYBRID SEARCH SHAREPOINT · GDRIVE · CONFLUENCE OCR · RBAC · AUDIT LOGS LOCAL LLM · OLLAMA
INCLUDED IN ALL PLANS
Optional modules — combine freely
MODULE 01
🗂️
Operations Hub
Tasks · Kanban · Projects · Milestones
Full project and task management with kanban boards, milestone tracking, and deadline control — all integrated with the RAG engine so your team can query project status in natural language.
  • Kanban board with custom columns
  • Milestone and deadline tracking
  • AI queries on task and project status
  • Cross-module context pinning
+ add-on per plan
MODULE 02
📊
Excel Copilot
Natural language → formulas, pivots, charts
Engineers without advanced Excel skills get any result they need. Describe what you want in plain language — the AI generates the formula, macro, or pivot table and explains the logic step by step.
  • Natural language to formula / macro
  • Pivot table and chart generation
  • Step-by-step explanation of logic
  • Works with existing workbooks
+ add-on per plan
MODULE 03
📖
Wiki & Knowledge Base
Procedures · Checklists · Version history
Built-in wiki with Markdown editor, version history, templates and page types — procedures, checklists, meeting notes. All pages are RAG-indexed automatically and searchable alongside your uploaded documents.
  • Markdown editor with templates
  • Full version history and diff view
  • Auto-indexed in RAG on save
  • Procedures, checklists, meeting notes
+ add-on per plan
MODULE 04
🏗️
Incidents & Asset Registry
Assets · Incidents · Regulatory deadlines
Register and track physical assets — elevators, equipment, installations. Link incidents to assets, track regulatory response deadlines, and auto-generate AI-written incident reports directly from task data.
  • Asset registry with status tracking
  • Incident linked to asset and regulation
  • AI-generated incident report (one click)
  • Regulatory response deadline alerts
+ add-on per plan
MODULE 05
🤖
AI Agent Mode
Autonomous multi-step task execution
Give the AI a complex instruction and it executes a multi-step plan autonomously — searching documents, creating tasks, generating reports, and updating assets in a single ReAct loop. Everything runs locally.
  • ReAct loop with 8 tool executors
  • Real-time SSE step streaming
  • Cross-module autonomous actions
  • Fully local — no external API
+ add-on per plan
MODULE 06
📈
Analytics & ROI Dashboard
Usage · Hours saved · Executive reports
Track query volume, response times (p50/p95/p99), hours saved per team, and recovered engineering time. One-click executive PDF report for quarterly reviews and ROI justification.
  • Query volume and latency tracking
  • Hours saved per team per week
  • One-click executive PDF report
  • ROI calculator for management
+ add-on per plan
Example: a 3-module stack for a maintenance company
Core + Incidents & Assets + Wiki — everything needed to query regulations, manage assets, track incidents with automated reports, and build internal procedures. No kanban, no analytics, no Excel. Pay only for what your team actually uses.
🧠
Core
always included
🏗️
Incidents & Assets
module 04
📖
Wiki
module 03
Your stack
configured at onboarding
All module combinations are supported · Modules can be added or removed at any time · Pricing scales with modules + seats
Process

From document to answer
in seconds.

No complex setup. No training required. Upload your documents, ask your question, get your answer.

01
Deploy on Your Server
IT runs docker-compose up. The full stack — AI models, vector database, app — starts automatically on your hardware.
02
Upload Documents
Upload PDFs, scanned documents or Excel files. EngiIntel deduplicates, processes and indexes them using local AI models.
03
Ask in Natural Language
Engineers type their question in plain language. The system checks the cache, then searches semantically across all indexed documents with cross-encoder reranking.
04
Get Cited Answers
Receive precise answers with document name and page number. Verify directly in the source in one click.
Industries

Built for regulated
environments.

Where documentation accuracy is a legal requirement, not a preference.

🛗
Elevator & Vertical Transport
BOE regulations, EN 81 series, inspection reports, maintenance schedules and modernisation projects — cited answers in seconds.
⚗️
Pharmaceutical
SOPs, validation protocols, regulatory submissions and GMP documentation — all queryable with full audit trail.
✈️
Aerospace & Defence
Maintenance manuals, airworthiness directives, technical orders and component specs — mission-critical accuracy.
Energy & Utilities
Plant documentation, safety procedures, IEC standards and inspection reports. Instant access during operations.
🏭
Manufacturing & Industry
Technical specs, CE documentation, machinery directives and quality manuals — queryable by any engineer.
🏗️
Construction & Infrastructure
Building codes, structural calculations, project specifications and inspection reports — all searchable from site.

Your data never
leaves your building.

EngiIntel was designed from day one for environments where data confidentiality is non-negotiable.

Zero cloud dependencyAI inference runs locally. No OpenAI, no Google, no external APIs required.
GDPR compliant by architectureData never crosses jurisdictional boundaries. Residency guaranteed.
Works fully offlineDeploy on your intranet. No internet required after initial setup.
IT has full controlStandard Docker deployment. Your team manages the server, backups and access.
# Architecture overview

[ Your Server ]
  ├── EngiIntel App (port 3000)
  ├── PostgreSQL + pgvector (port 5432)
  ├── Ollama LLM (port 11434)
  │    ├── mxbai-embed-large
  │    └── llama3.1:8b
  └── pgAdmin (port 5050)

─────────────────────────
✗ OpenAI API  not connected
✗ Google API   not connected
✗ Any cloud    not connected

✓ 100% local inference
✓ Zero data egress
✓ Offline capable
Pricing

Scales with your team.
Not with your queries.

Core is always included. Add only the modules your team actually uses — the calculator below shows your exact price.

Configure your plan
Select seats and modules — price updates in real time
users
🧠 Core (RAG)
🗂️ Operations
📊 Excel Copilot
📖 Wiki
🏗️ Incidents
🤖 AI Agent
📈 Analytics
Core · per seat €8/mo
× seats 5
Total / month
billed monthly · cancel anytime
€40
Request this plan →
On-premise · no per-query costs · Docker install
Starter
8
per seat / month · Core only
  • Core module (RAG + connectors + OCR)
  • From 1 seat — no minimum
  • Unlimited documents & AI queries
  • Add any module: +€4/seat each
  • On-premise — your infrastructure
  • No SSO / Entra ID
Start Free Trial
Enterprise
32
per seat / month · all 6 modules
  • Core + all 6 modules (+€4 each)
  • Volume discount from 15 seats (−10%)
  • Microsoft Entra ID SSO
  • Dedicated SLA + 24h support
  • Custom onboarding + ERP integration
  • Multi-site / multi-department
Request Enterprise Demo

All plans: on-premise · Core €8/seat · each module +€4/seat · volume discounts from 15 seats · no per-query costs · cancel anytime

Roadmap

Built in public.
Shipping fast.

EngiIntel is production-ready. All core features are live. Active development continues on enterprise scale and deeper integrations.

Foundation
✓ Complete
DockerLocal AIOCRRAG PipelineJWT AuthAudit LogsAdmin PanelSelf-hosted DB
Connectors & Security
✓ Complete
SharePoint SyncGoogle DriveConfluenceEntra ID SSOPII DetectionMultilingualRBACFolder System
Platform & Billing
✓ Complete
Multi-tenant SaaSStripe BillingROI DashboardOnboarding WizardPublic REST APIWebhooksGDPR ErasureSOC 2 Export
Answer Quality
✓ Complete
Confidence ScoringReAct Agent LoopQuery ExpansionCross-Encoder RerankingChunk Summaries
Performance & Scale
✓ Complete
Query CacheDoc DeduplicationTop-20 RerankLatency p50/p95/p99Lazy Loading
Wiki & Incidents
✓ Complete
Wiki EngineVersion HistoryAsset RegistryIncident ReportsAI Report GenRAG Auto-Index
Context Intelligence
✓ Complete
Passive ContextPinned PanelCross-Module AIFull ResolutionWiki Sources in RAG
AI Agent Mode
✓ Complete
ReAct Loop8 Tool ExecutorsSSE StreamingStep RenderingAutonomous Tasks
Enterprise Scale
Q3 2026
SAML 2.0 / SCIMIP AllowlistingvLLM ScaleSIEM ExportCustom SLA
FAQ

Questions IT always asks.

Straight answers to the questions that come up in every security review.

Where is our data stored?+
100% on your own server. EngiIntel stores all documents, embeddings and query logs in a PostgreSQL database running in Docker on your hardware. No data is ever sent to any external service.
Does it work without internet?+
Yes. After initial setup and model download, EngiIntel operates entirely offline. It can be deployed on an air-gapped intranet with no outbound internet access required.
What hardware do we need?+
For small teams (1–10 users): a modern workstation with 32GB RAM and a mid-range GPU (8GB+ VRAM) runs comfortably. For larger deployments a server with GPU is recommended. Runs on Windows, Linux and macOS via Docker.
Is it GDPR compliant?+
Yes, by architecture. Since all data stays on your infrastructure, there is no cross-border data transfer. Includes a GDPR right-to-erasure endpoint, PII detection, and full audit trail export.
What happens if the server goes down?+
EngiIntel is a standard Docker application. Your IT team manages backups, redundancy and uptime using your existing infrastructure procedures. Enterprise plans include deployment assistance and SLA documentation.
Get Started

Ready to deploy AI
on your infrastructure?

Book a demo and we'll show you EngiIntel running live on a real document set from your industry — no slides, no mockups.

📍Barcelona, Spain · Remote demos available
Response within 24 hours

No commitment required. We'll reach out within 24h.