01
Trust it
Quality rules and signals are monitored continuously, so teams can see whether governed data meets agreed expectations.
What it does
The same source of truth, for your team and your AI.
Four platform capabilities work together across reports, datasets and metrics, whether a person or an AI assistant is asking.
02
Find it
Search every report, dataset and definition in plain business language, not table and column names.
03
Understand it
One glossary, harmonised across departments, so finance and operations argue about the business, not the number.
04
See it in context
Governed context is available through search, supported BI extensions, APIs and AI assistants.
Open by design
One governed layer, available through every useful surface.
Search is one interface. The BI extension is another. APIs and MCP let your own applications, chatbots and agents retrieve the same approved context.
Single sign-on
Use SSO so employees enter dScribe through the identity setup your organization already manages.
Open APIs
Read and connect governed context through APIs and webhooks, without locking every workflow into one interface.
MCP for assistants
Connect chatbots and agents through dScribe's MCP server so they can retrieve the same approved business context people find in search.
Metadata loader
The catalog starts building itself.
Connect your source systems and dScribe crawls their metadata automatically: structures, names, descriptions, lineage and usage. Your team does not have to build the inventory by hand.
Scheduled refreshes keep the catalog fresh as schemas, reports and systems change. People still own the judgment calls: ownership, policy and business meaning.
Metadata loaderConnected sources
SyncedMicrosoft Fabric
1,248 assets
Metadata refreshed automatically
Databricks
386 datasets
Metadata refreshed automatically
Snowflake
214 tables
Metadata refreshed automatically
Power BI
172 reports
Metadata refreshed automatically
Governed suggestions
Let the catalog find the work. Let people make the decision.
The catalog does the scanning work: it looks across the metadata it collects and surfaces likely relationships, duplicates, missing links and columns that need documentation. That is the self-driving part. Owners still review each suggestion before it becomes governed context, so the catalog stays current without manual audits.
Suggestions for review
4 openRelationship suggested
Review relationshipCustomer orders may relate to customer accounts
customer_orders.customer_id matches customer_accounts.customer_id across recent metadata scans.
Documentation suggested
Review descriptions12 columns in sales.orders lack descriptions
Adding these by hand would take hours. dScribe drafted descriptions from usage, names and lineage for the owner to review.
Duplicate detected
Review duplicateMonthly revenue appears twice in Finance
Two terms share the same owner, calculation and connected report set.
Link suggested
Review linkGross margin should be linked to finance.margin_pct
Usage and naming indicate that this column implements the approved glossary definition.
With the tools you already run.
Power BITableauMicrosoft FabricDatabricksSnowflakeSAPDynamics 365
Where to start
Start with the problem that hurts most.
Each card below is a different entry point. They all run on the same context layer, so the owners, definitions and rules you establish in one area carry into the others.
Data governance
Stop postponing decisions because no clear owner is defined.
One person owns each report and dataset, and everyone can see who. No mailbox, no wiki.
Explore data governanceBusiness glossary
Stop settling arguments in meetings. Settle them in the glossary.
Every metric gets one agreed definition, shown exactly where the question comes up.
Explore the business glossaryData contracts & quality
Stop learning about data problems only when someone reports them.
A contract and a quality check between producers and consumers, built on the open ODCS standard.
Explore data contracts & qualitySemantic layer for AI
Don't let your AI initiative outrun your governance.
Give your AI agents the same governed definitions your BI reports already trust, not a guess.
Explore the semantic layer for AIWhat changes once the context layer is in.
<15 sec
to an answer that used to mean a ticket and a day's wait
3+ hrs
back every week, per person on the data team
100%
increase in data-informed decisions once trust is established
Already working this way
Talk through where governed context fits.
A short conversation about your stack, users and AI initiatives.
What people usually ask at this point.
Do we need to migrate our data into dScribe?
No. dScribe reads directly from Power BI, Tableau, your warehouse and your knowledge systems, in place. Nothing is copied out or duplicated.
How is this different from a data catalog?
dScribe is a self-driving data catalog. It crawls metadata automatically and suggests relationships, definition links and duplicates for people to review. Your team spends less time maintaining an inventory and more time approving trusted context that grounds AI use cases.
Does dScribe make changes automatically?
dScribe automates discovery and prepares suggestions, but governed changes remain reviewable. Owners decide whether to accept a relationship, merge a duplicate or link a definition before it becomes approved context.
How does this help us prepare data for AI?
The platform shows which assets already have approved meaning, ownership and relationships, and which still need attention. Teams can focus their governance effort on the context their chatbots and agents will actually use.