Module · Index

All your data. Every format. One knowledge layer.

Index understands structured and unstructured data — spreadsheets and contracts, scans and 3D models — and turns them into one searchable, permission-aware source of truth. For people who search. For AI that answers. For systems that need clean master data.

1
knowledge layer for every module and system
All
formats — from Excel to BIM models
100%
permission-aware — every answer, every hit
index · knowledge
Sources
Datasheet_X200.pdf
PDF · scan
Pricelist_Q3.xlsx
Excel
Supplier · Update
Email
Building_B.ifc
BIM / 3D
Golden Record
Quality 98%
Article X200-4417
EAN90083901441733 sources
Price€ 1,284.00by minute
Stock412 pcsERP · live
Datasheetv4 · 12 pagesindexed
Search semantic or structured…
Structured & unstructured

A document is more than text. Index understands all of it.

A scanned contract contains text, tables, stamps and signatures. A price list contains structure. An email attachment contains both. Index breaks every source down into what it really is: text becomes semantically searchable, tables become SQL-queryable, images get described, entities get recognized. A pile of files becomes a knowledge layer.

  • OCR for scans and photo PDFs — even old archives become searchable
  • Tables from PDFs and Excel as real data: filter, aggregate, query
  • Images and diagrams described automatically and made findable
  • 3D building models (IFC/BIM): elements and properties as queryable data
  • New content is searchable in seconds — processing runs in the background
index · ingest
Incoming
Contract_2481.pdf
Scanned · 24 pages
OCR · text recognition
Understood as
Text
42 chunks · semantic search
Tables
3 tables · SQL-queryable
Images
5 images · auto-described
Entities
supplier · article · project
Searchable in seconds — text, tables, images
Understands, among others
PDF & scansWordExcelPowerPointEmailsHTMLCSVImagesIFC / BIMDatabase tables
Knowledge graph

Index doesn't just know content. It knows connections.

During ingestion, Index recognizes entities — suppliers, articles, projects, people — and links them across sources. The article in the datasheet is the same one in the order and in the contract. These connections make search more precise and AI answers better grounded: context comes not from one document, but from all of them.

  • Entities and relationships extracted automatically — no manual tagging
  • Links across formats: contract ↔ article ↔ order ↔ project
  • Search with context: related facts and neighboring entities in every hit
  • The foundation that lets Assistants answer precisely, with sources
index · graph
TechParts GmbH
supplier
Datasheet_X200.pdf
document
Article X200-4417
12 relations · 4 sources
Order 10382
order
Project Nord
project
Entities & relations · extracted automaticallyacross all formats
Master data management

Golden records: one clean truth from many sources.

The same article in your ERP, the supplier catalog and the webshop — with three different prices and two spellings. Index consolidates sources into a golden record: deduplicated, validated, with per-field provenance. Every change is versioned, every rule documented.

  • Consolidation and deduplication across source systems
  • Per-field provenance: which value comes from where — and why it wins
  • Data quality rules and completeness scoring per record
  • Versioning and change history — every state traceable
  • Documents, images and datasheets attached to the record — not next to it
index · master data
Golden Record
Quality 98%
Article X200-4417
EAN90083901441733 sources
NameRack unit 42UERP · master
Price€ 1,284.00API · live
CategoryRacks / 42Urule
Datasheetv4 · 12 pagesindexed
History
v14
Price updated
Pricing API
v13
Datasheet v4 linked
Index
v12
Duplicate merged
2 sources
v11
Category changed
awaiting approval
Every change versioned · full audit trail
Governance

Six features that keep your data under control.

Who sees what, where does a value come from, and who changed it? Index has built-in answers to these questions — from permissions and versioning to tenant isolation.

Permissions on every access

Every search and every query respects your permission model — inherited from your identity provider, enforced down to the individual search hit.

Versioning & audit

Every change to content and master data is versioned and attributed — complete audit trail included.

Per-field provenance

Every value carries its source: which system, which document, which rule. No black-box data.

Multi-tenant

Strict isolation per tenant — for group structures, subsidiaries or your partners' customer setups.

Quality scoring

Completeness and rule compliance visible per record — data quality becomes measurable instead of assumed.

Multiple indexes

Separate indexes per department, project or data domain — each with its own lifecycle and its own permissions.

For humans & for AI

Index once. Use everywhere.

The same knowledge layer answers an employee's search, a question to an assistant and a third-party system's API call — with the same data and the same permissions.

For humans
  • Semantic search in natural language — across all formats
  • Structured queries on tables: filter, aggregate, export
  • Every hit with source and context — down to the original document
  • Browse, review and maintain golden records
For AI & systems
  • Retrieval foundation for Assistants: answers with evidence, not hallucination
  • MCP interface: external AI agents use Index too — with permissions enforced
  • REST API for third-party systems: webshop, portal, custom builds
  • Flows read and write directly — Index is the platform's hub

Permissions apply everywhere equally: what a user can't see, their AI can't see either.

FAQ

Questions about Index.

How is Index different from a classic PIM or MDM solution?
Classic MDM systems manage structured records. Index additionally manages unstructured content — contracts, scans, emails, images, even BIM models — and connects the two: the datasheet is attached to the article, the contract to the supplier. The same layer doubles as the retrieval foundation for AI.
How does our existing data get into Index?
Through Flows: connections to ERP, file shares, email inboxes and 200+ other systems. Existing documents are bulk-imported — OCR for scans included. New content flows in automatically from then on.
Does Index replace our ERP or our DMS?
No. Index sits as a knowledge layer on top of your systems: it reads from them, consolidates and makes everything searchable. Systems of record stay in charge — Index makes their data usable, including for AI.
How does Index make sure AI answers are correct?
AI answers only from indexed content, and every answer carries its sources. The knowledge graph provides context across document boundaries, and permissions filter before the answer — not after.
Can we start with one area instead of all our data?
Yes — and that's the usual path: one index for one domain, such as article master data or project documents, then expand from there. Multiple indexes with separate permissions are built in.

Show us your messiest data estate.

30 minutes. We'll show you what it looks like as a knowledge layer — searchable, cleansed, ready for AI.