SCENTRY
Agentic ingestion · Multi-modal search

Everything is already in the documents.

Point Scentry at thousands of PDFs, scanned files, and spreadsheets. A chain of agents reads every page on its own, resolves the people, companies, and money between them, and assembles a high-quality knowledge graph — then lets you search it by meaning, by relationship, and by exact value at once. No schema, no pipeline, no analyst in the loop. All inside your environment.

Read
docs + tables in
Extract
entities + links
Resolve
merge duplicates
Connect
query the graph
undisclosed common owner
Documents47 filesDatasets3 sheets → 1possibly same · 87%AGENTWho controlsBalticInfrastructure?→ tracing ownersMinistry of TransportBaltic InfrastructureMeridian HoldingsInfra Baltic OÜContract · €4.2MTallinnViktor Kovalenko
1,389 entities · 4,102 relationsresolved
PersonOrganizationLocationEventAccount
Why it's different

Every answer traces back to the page it came from.

Scentry is built for evidence you can stand behind — connections made explicit, sources kept intact, and nothing taken on faith.

01
Everything is an entity

People, organizations, transactions and the relationships between them are first-class objects — not strings trapped in a PDF or a spreadsheet cell.

02
Provenance on every claim

Nothing is asserted without a source. Each fact links straight back to the exact page, row or cell it was drawn from — so any answer is yours to verify.

03
Faithful to the original

Evidence stays in its native language and form. Nothing is silently translated, normalized away, or quietly dropped on the way in.

What it resolves
Five entity types, one resolved graph.
Person
owners · directors
Organization
companies · authorities
Location
registered addresses
Event
tenders · contracts · payments
Financial account
shared-account pivots

Every entity is a projection over the mentions that named it — each one tracing back to the exact page, row, or cell it came from.

In the app

Every workspace, at a glance.

Projects, documents, datasets, entities and flagged anomalies — all tracked as the agents work, all inside your environment.

app.scentry.ai/dashboard
Dashboard Search New Project
Projects
3
3 active
Documents
82
across all projects
Entities
1,389
892 high confidence
Anomalies
25
14 high severity
Active Projects
NL Procurement 202447 docs · 3 datasets · 892 entities · 142hactive
Baltic Shell Network23 docs · 1 dataset · 341 entities · 81dactive
Cyprus Vendor Audit12 docs · 2 datasets · 156 entities · 33dcomplete
Recent Activity
47 entities extracted from Vendor_Registry.xlsx2h ago
Anomaly: circular ownership in Baltic Shell Network4h ago
Contract_CT-2024-0881.pdf processed (14 pages)6h ago
On your terms
Scentry runs where your evidence already lives — and what it reads stays there.
Request a briefing