No modules to license one by one. Every capability below ships in MDM Studio, works on every supported engine, and is managed from a single, modern web studio.
One canonical model served natively on SQL Server, PostgreSQL, MySQL and Oracle. Switch engines per session — no separate installs, no forked models.
Run many isolated organisations on one deployment. Every model, source, rule and golden record is tenant-scoped, with per-tenant branding and cross-tenant membership.
Users sign in with their AD / LDAP credentials or a local password — chosen per user. Each tenant configures its own directory with an encrypted service account and a connection test.
Classify domains, entities, attributes and records from Public to Restricted. Per-user clearance grants control who sees what; restricted records are hidden outright, not just masked.
Access groups decide which records each person sees — scoped by domain, hierarchy tier, value-hierarchy level or a column value, and composed across their groups. One rule filters every surface, so counts, quality scores, queues and exports all agree. Domains are governed only once armed, and a refusal is a named answer, never an empty grid.
Model multi-level hierarchies with per-tier dimensions, labels and data-quality scoring — so quality is measured where it actually matters in your structure. Tier names are set on the model and overridden per domain, so a Sales domain and a Finance domain can call the same level different things; every tier binding an entity carries is edited in one place on the entity form.
Deployment models with clone, XML import/export and checkout. Domains, entities, a governed data dictionary and an ERD studio with visual and translated exports — all AI-enriched, from suggested classifications to golden-record scoring, with live previews from your real sources.
Multiple connection types — SQL Server, Azure SQL, Synapse, PostgreSQL, MySQL, Oracle, Teradata, CSV, Excel, FTP and REST APIs — in one global registry.
A guided, stage-aware pipeline: mapping sets that build themselves stage to stage, drag-to-create jobs with auto-allocation into groups and packages, delta-load guidance, and one operations console for every run.
Scored match rulesets with human review and phonetic and fuzzy comparators, per-attribute survivorship strategies, and golden records that explain themselves — evidence-based confidence, full cross-reference lineage and a superseded-not-deleted lifecycle.
Compare two systems of record before either is trusted. Match by golden records, or by table — pick the tables, the record key on each side (a column, or a key derived from several: concatenate, split, digits, substring), and the fields to line up. Preview the key first: how many rows, how many unique keys, what the other side joins to. Distinct collapses exact copies; repeated keys report as within-source conflicts instead of stopping the run. An advisor ranks key candidates from the data and narrates why.
AI-suggested validation rules backed by a curated pattern library, plain-language operators, scorecards and snapshots — and a closed remediation loop: typed resolutions with automatic re-validation, auto-remediation and honest campaign progress. Group by rule when a scorecard should read as a list of rules rather than a list of records.
Governed code sets with immutable releases, crosswalks between source vocabularies, and a read-only serving API with revocable service tokens.
Change requests through review and approval, governance policies, tamper-evident audit trails with per-event detail, model checkout — and a review bypass that visibly marks a governance-suspended model, plus a strict pipeline mode that makes Survive and Publish earn their run. A request names the objects it touches by browsing your model, not by typing a name, so every targeted item carries a real id and a live label; linking, wiring and removing one are all in the event trail. And the people who reviewed, approved and implemented it can sign their part — an attestation bound to a fingerprint of the record’s facts, so editing the record afterwards marks those signatures stale instead of carrying them forward.
Push mastered values back to source systems under eligibility rules, batching, approval gates and post-apply verification.
A full report studio inside the hub: seven visual types, eleven chart types, calculated fields compiled to SQL, pivots with drill, show-as windows (% of total, % of a grouping, running totals, rank), a band editor that can suggest bands from the data's own distribution, column tools that draw one rule set as an indicator, heat bar, badge, range or delta over numbers, text or dates, dashboards whose filters unify across sources, saved views, scheduled subscriptions and server-side export. See what it does →
Follow one record through everything it touches. Walk its relationships tab by tab, pivot onto any related row with a breadcrumb behind you, and search a golden key through those relationships — a Sales Order key finds the Customer that owns it. Every tab carries a row count before you open it, and a search that could not reach a relationship says which one and why. See it →
Number, date and text formats governed once on the attribute in the data dictionary and honoured by every surface — grids, chart labels, axis ticks, KPI cards. Display only: filters, copy, CSV, generated SQL and every API payload keep the raw value.
The whole product flips theme from a preference stored on your account and painted before the first pixel. Contrast is swept across every page in both themes as part of the build — a regression fails the build rather than shipping.
Shared defaults that resolve from your own preference, then the model default, then the system — and tell you which tier answered. The shared default is a governed, audited change; your own is yours to set.
Role-based access control down to page and action, user management, idle-session policy, auto-provisioned hub databases, per-hub health diagnostics and one-click hub housekeeping.
Move a model's configuration from development to UAT to production over a channel the receiving hub controls — it issues the token, decides what it accepts, and can refuse. Always dry-run first: the receiving hub checks column compatibility across builds, refuses to land on a model somebody has checked out, and archives everything it currently holds before it changes anything. Both hubs log every attempt, including the refusals.
MDM Studio evolves in rapid, versioned releases — every build is tracked, audited and visible right inside the product.
Tell us what matters most — matching, governance, multi-tenancy, reference data — and we'll tailor a live walkthrough to it.