Platform · Data Catalog

Your data, catalogued and understood in minutes, not months.

Connect a database, document store, or PDF corpus and LazyFox discovers every table, collection, and document, infers every relationship, and generates semantic metadata from the data itself. No manual tagging. No documentation sprints. No consultants.

See a demo → Explore capabilities
8 tables indexed, described, and related: under 2 minutes·
Soft relationships inferred where no foreign key exists, confidence-scored and auditable·
AI-generated description for every column: meaning, data type, possible values, date format·
Schema drift tracked automatically: reports flagged obsolete when underlying tables change·
Read-only scanning: zero writes to your production database, ever·
Works with Azure SQL, PostgreSQL, MySQL, MongoDB, Databricks and more·
PDFs and document stores indexed and opened to structured AI access·
8 tables indexed, described, and related: under 2 minutes·
Soft relationships inferred where no foreign key exists, confidence-scored and auditable·
AI-generated description for every column: meaning, data type, possible values, date format·
Schema drift tracked automatically: reports flagged obsolete when underlying tables change·
Read-only scanning: zero writes to your production database, ever·
Works with Azure SQL, PostgreSQL, MySQL, MongoDB, Databricks and more·
PDFs and document stores indexed and opened to structured AI access·

Not a documentation project.
A living catalog.

Traditional data catalogs rest on a false premise: that someone has time to manually tag every table, write column descriptions, and keep them current after every release. In practice, that work never gets done. IDC puts less than 1% of enterprise data in active AI use today, most of it unstructured and invisible to the tools meant to query it.

LazyFox generates your catalog the moment you connect a source: a relational database, a document store like MongoDB, or a folder of PDFs. It scans schema and content, infers what each asset means and how it relates to others, and builds a versioned semantic layer that stays current as your data evolves. Your team doesn't write a word of documentation.

Traditional Data Catalog LazyFox Data Catalog
Manual tagging by data stewards
AI-generated from schema and data samples
6–12 months to implement, before any value
Running within minutes of connecting a source
Stale after every release
Re-indexes automatically, surfaces drift
Hard foreign keys only
Soft relationships inferred and confidence-scored
Siloed tool, rarely consulted
Foundation for every AI query in LazyFox
Expensive implementation engagement
One connection, one configuration step

Everything the indexing job does —
and what happens after.

Six capabilities that run automatically when you connect a source, and keep running as your data changes.

🔍

Auto-Discovery

Every table, view, column, index, partition, and clustering key found and catalogued automatically. Table sizes, row counts, and analytical profiles are included so you know what you're working with before you write a single query.

Schema · Profiling · Instant

AI Enrichment

LazyFox samples each table and runs AI enrichment over every column, generating plain-language descriptions, inferring data types, identifying possible values, and recognising date formats. No human writes a word of it.

AI-generated · Column-level
🔗

Relationship Inference

Hard foreign keys are just the start. LazyFox uses naming patterns, data samples, and LLM reasoning to detect soft joins (relationships your database never declared but that exist in practice). Each one comes with a confidence score and can be reviewed or overridden.

Hard + Soft · Confidence-scored
🗂️

Logical Grouping

Tables are grouped into logical domains: customer data, transaction data, product catalogue, and so on. When someone asks a business question, LazyFox routes it to the right cluster immediately rather than scanning every table from scratch.

Domain-aware · Query routing
📅

Schema Drift & Versioning

Every change to the semantic layer is versioned with a full diff. When a table's meaning shifts (new columns added, descriptions updated, relationships changed), reports and queries built on the previous version are flagged as potentially obsolete.

Versioned · Diff · Report flags
🛡️

Audit Trail & Guardrails

Every indexing run, edit, and query is logged. Access is enforced at the column level, scanning is strictly read-only, and the same governance controls extend to unstructured sources: PDFs and document stores receive the same audit trail, access policies, and lineage tracking as any relational table.

Read-only · Column-level · Logged
LazyFox · Data Catalog · Indexing run
transactions
14 cols · 2.4M rows Done
customers
22 cols · 180K rows Done
accounts
18 cols · 180K rows Enriching
statements
9 cols · 640K rows Queued
products
11 cols · 312 rows Queued
Enriched · transactions
Table Records all debit and credit movements on customer accounts, including card transactions, transfers, and fee charges.
amount Transaction value in EUR (cents). Negative = debit. Range: –50,000 to +100,000.
status Possible values: PENDING · SETTLED · REVERSED · FAILED
Soft relation detected → statements.statement_id — appears in 99.8% of rows (no FK declared)

One job. Everything your team needed to know about the data, generated automatically.

When you connect a source, LazyFox runs a single indexing job. It doesn't need a schema diagram, a data dictionary, or a briefing from the people who built it. It figures it out from the data itself.

The result is a semantic layer that covers every asset in your data source: every table, collection, or document, what it contains, what each field or section means, how it relates to other assets, and how confident those relationships are. It's built once, updated automatically, and available to every query from that point on.

  • Discovers hard keys from schema and soft joins from data samples and LLM reasoning
  • Runs off-hours on a configurable schedule, with minimal footprint on production
  • High-activity tables re-indexed daily; dormant ones less frequently
  • Works across relational databases, document stores like MongoDB, and unstructured files including PDFs
  • Builds a knowledge graph and vector index used by every downstream AI query

Your catalog doesn't go stale after the next release.

Every change to a table's semantics is tracked as a version. When a column changes meaning, a relationship is dropped, or a new field appears, LazyFox generates a diff and surfaces it before that change quietly breaks a report or returns a wrong answer.

Any report or dashboard built against a previous version of the semantic layer is automatically flagged. Your team sees exactly what changed and can decide whether to regenerate or accept the existing output.

This matters most in regulated environments where data lineage must be documented. A stale definition has audit consequences, not just analytic ones.

Semantic diff · accounts · v1 → v2
accounts · description
v1 v2 · current
table: accounts
description: "Stores customer account records."
+description: "Stores current and savings account records for retail banking customers. Excludes corporate accounts held in accounts_business."
 
column: balance
type: integer
+type: decimal(18,4) — updated after migration v2.14
 
+soft_relation: accounts_business.account_id (NEW · 94% confidence)
3 reports affected by this change
Monthly Balance Summary, Q2 Regulatory Report, and CFO Dashboard were generated against v1 semantics. Review recommended before next distribution.

One catalog.
Five different problems solved.

The Data Catalog is built for people who have been living with undocumented data, and for the AI that needs to query it reliably.

⚙️
Data & Infrastructure Engineers
Understand the schema they inherited without chasing the vendor for documentation they'll never get. Know which tables own which data and how they connect.
Discovery
🏛️
Data Warehouse Teams
Know exactly which source tables to pull from and how to join them correctly — before building a pipeline, not after debugging one.
Accuracy
📊
Business Analysts & Chief of Staff
Ask questions in plain language without manually joining 12 tables. The catalog makes data self-serve for people who are not engineers and shouldn't need to be.
Self-service
⚖️
Compliance & Risk
Documented data lineage, column-level access controls, and a full audit trail — the foundation of BCBS 239, DORA, and GDPR data governance requirements.
Governance
🤖
AI & Analytics Agents
Every LazyFox AI query runs against the catalog — not raw schema. The semantic layer is what makes answers deterministic and trustworthy rather than probabilistic guesses.
Reliability

Works across the sources your data already lives in.

LazyFox connects to relational databases, document stores, and unstructured file sources used in financial services and enterprise environments: on-premise, cloud-managed, or hybrid. That includes PDFs, which LazyFox opens to structured AI access alongside your transactional and operational data.

See all integrations →
Azure SQL Managed Instance SQL Server PostgreSQL MySQL / MariaDB MongoDB Databricks Snowflake BigQuery Redshift Oracle PDF / Documents

See your data understood in 5 minutes.

We'll connect a live data source in the demo (a database, document store, or PDF corpus) and walk you through what the catalog generates: tables, collections, enrichment, soft relationships, and all.

About this page

What is the LazyFox Data Catalog?

The LazyFox Data Catalog automatically discovers, documents, and semantically enriches enterprise data assets. Connect any data source (a relational database, a document store like MongoDB, or files such as PDFs) and it scans every table, collection, and document, infers relationships, and generates business-readable metadata from the data itself, with no manual tagging, documentation sprints, or consultant-led projects needed. The catalog is live within minutes of connection.

The catalog stays current automatically. When schemas change or new tables are added, it updates. When business users change a metric definition, it reflects that too. There is no re-indexing job to schedule and no documentation sprint to keep things accurate.

LazyFox also covers unstructured data governance: PDFs, document stores, and non-relational sources are catalogued, access-controlled, and lineage-tracked with the same rigor as structured databases. Organizations in regulated industries use this to close the compliance gap that typically exists between their relational data and everything else.