Building a data lakehouse
is simple now.
The DataSimple mark is a house built from rows of data, standing on a lake: every system you run, brought together under one governed roof, on open storage in your own cloud.
DataSimple Lakehouse: connect every system you run, get one trusted copy of your data in your own cloud, and ask it questions in plain words.
Transform data for AI readiness. Fast.
Real numbers from a DataSimple load: one table, one run, straight from the source into Delta Lake and Iceberg.
Written to the lakehouse in seconds.
One 150-million-row table copied from the source, written as Parquet, Delta Lake and Iceberg, and applied to Bronze, in a single run.
Big loads, small footprint. DataSimple runs on a small 2 vCPU, 8 GB application server. Large tables are processed in parallel by DataSimple's breakthrough load process, which scales on demand and steps away when the load is done.
Measured on a DataSimple installation (run #25 on 26 Oct 2026: 150M rows, one table). Speed depends on the source, network and compute size.
Data that is ready for AI the moment it lands.
AI is only as good as the data under it. DataSimple turns raw source data into clean, governed, documented tables that people and AI can trust.
DataSimple turns your business data into something simple to read and understand.
DataSimple is a simple application for any business that wants to see its data clearly. It connects to as many source systems as you have and brings every one of them into Delta Lake. Anyone can start using it without prior knowledge of data models.
Pick your formats, and DataSimple builds the rest: replication with change capture, Bronze, Silver and Gold tables, security, a conversational AI you can simply talk with, and forecasting and planning on top. Your data family, end to end, in one application in your own cloud.
No data complexity
DataSimple takes the complexity out of your business data. Pick your formats, and it builds the rest.
Easy to understand your business data
Ask Everything in your own words, or just speak to Sarah. Your data answers back with the numbers, the reasons and what to do next.
Simple to run any business data
Whatever the challenge, DataSimple handles it, so you can focus on running your business.
Four steps from source system to answers.
Connect, replicate and store, model, and ask. Pick a step to see what it looks like in DataSimple.
Connect unlimited sources. Land them all in Delta Lake.
There is no limit on how many source systems you connect. Add databases, files, APIs, Kafka, or even a backup file, and bring each one into Delta Lake (or Iceberg and Parquet). Before you load anything, you can query the source live from a notebook and join it with your lakehouse tables.
- Unlimited source systems: SQL Server, Oracle, Db2 and COBOL systems, SAP ECC, Salesforce, APIs, files and more
- SAP ECC and Salesforce replicate as easily as any database
- A backup (.bak or dump) is restored and loaded for you
- Live queries across several sources in one notebook
SELECT c.region, SUM(o.order_total) AS open_value FROM src.erp_sqlserver.dbo.sales_orders o JOIN gold.dim_customer c USING (customer_id) WHERE o.status = 'OPEN' GROUP BY c.region;
Copy once. Then only what changed. Stored the way you choose.
DataBridge takes the first full copy, then brings only new, changed and deleted rows. Pick Parquet, Delta Lake or Iceberg, or all three, and every table lands in Bronze in your own cloud storage.
- Changes by watermark, by key and row hash, or from the database change log
- A load registry: nothing missed, nothing doubled
- Fast parallel extraction, with a bulk path for very large tables
Shape Silver and Gold in SQL, Python, or a sentence.
Apply changes to Silver with full history in one step, then build Gold facts and dimensions with notebooks, dbt and scheduled Dataflow pipelines. Or point at columns and say the rule with the Kinetic Pen.
- SCD Type 1 and Type 2 history, deletes kept
- Notebooks, dbt and Dataflow on a schedule
- Kinetic Pen: rules from pointing and plain words
# Current values plus full history, from the Bronze change log ds.apply_changes("silver.customers", "bronze.customers", keys=["customer_id"], sequence_by="_source_ts", stored_as_scd_type=2) # A Gold fact, refreshed on a schedule sales = read_table("silver.sales_orders") save_table(sales.groupBy("region", "week").sum("net_sales"), "gold.fact_sales_weekly")
Ask in plain words. Get the report.
Everything reads your Gold tables, with your permissions, and answers with tiles, charts, maps and a forecast with its likely range. Drill in place, ask follow-ups, or talk to Sarah.
- Forecasts tested on your own history before they are shown
- What-if scenarios from your Gold data
- Schedule a report and it refreshes itself
dbo.sales_ordersdonedbo.customersdonedbo.order_linescopyingdbo.productscopyingA lakehouse used to be a project. Now it is a product.
A lakehouse joins the open storage of a data lake with the structure of a warehouse. Building one used to mean stitching tools together and writing the plumbing yourself.
DataSimple gives you the whole thing, ready to use: one application, one copy of the data in open formats, in your own cloud.
Simple to use. Built for the enterprise.
DataSimple is a complete product with security, compliance, multi-cloud storage and catalog governance built in, and a clear process from raw data to decisions.
The platform, in four 3D drawings.
Hover or click any part to read what it does, press play to follow the steps, or switch to the flat image.
DataSimple has two planes: the control plane, the DataSimple services you work with (screens, governance and orchestration), and the data plane, the engines that do the work and the data lake that holds your data, in your own cloud storage. Hover a part of the drawing to read what it does.
People and scripts enter through one control plane; every table is named by the catalog; work runs in your own Spark session on your own cloud storage; outside readers get Gold over Delta Sharing.
The work always follows the same path: extract from the source, land a protected copy, keep every change in Bronze, transform in Silver, and serve the business from Gold. Press play, or click a stage to jump there.
Arrows show the direction data moves. Option 2, without DataBridge: a notebook reads the source and writes Bronze directly.
Ask for a business report or for platform work. No assistant has privileges of its own: it acts as you, and every change waits for a person.
Ask for a business report or for platform work. No assistant has privileges of its own; it acts as you, and every change waits for a person.
Every screen and every script goes through the same front door. Nothing in the browser or your scripts holds a password, key or engine address.
Every screen and every script goes through the same front door. Nothing in the browser or your scripts holds a password, key or engine address.
Six things you do not have to build.
01One application for the whole data lifecycle.Product
Replicate, store, model, govern, ask, forecast and plan in one place, instead of a stack of separate tools.
- Replication with CDC from unlimited sources
- Notebooks, dbt and Dataflow pipelines
- Everything and Sarah for answers
Why it matters: fewer tools to buy, connect and keep in sync, and one place to find any number.
02Secure by default, at every layer.Security
Data stays in your own cloud account. Every request is checked against who is asking and what they may see.
- Single sign-on and roles
- Row filters, column masks and PII masking
- Tenants fully separated: data never crosses
Why it matters: sensitive data such as salaries or customer details is protected the same way in SQL, Python, reports and voice.
03Evidence for every audit.Compliance
Every run, change, grant and approval is recorded, so you can show exactly what happened, when and by whom.
- Audit trail and history of every object
- Lineage from source to report
- Time travel and restore of earlier versions
Why it matters: answering an auditor or a privacy request takes minutes, not weeks.
04Your cloud, your choice.Multi-cloud
Store your lakehouse in the cloud you already use, in open formats any engine can read.
- Delta Lake, Apache Iceberg and Apache Parquet
- Move or add clouds without rewriting your data
- No proprietary storage, no lock-in
Why it matters: you keep control of cost, location and future choices.
05One catalog, modelled on Unity Catalog.Catalog and governance
Every table, view, grant, filter, mask, lineage edge and business meaning lives in one governed catalog.
- Three-level names: catalog.schema.table
- GRANT and REVOKE in SQL, information_schema
- Optional sync to an open-source Unity Catalog server
Why it matters: rules are set once and enforced everywhere, so governance is not a second project.
06A clear path from raw data to decisions.Process
Bronze, Silver and Gold layers give every team the same way of working, with checks and approvals built in.
- Data quality rules and certified tables
- Changes by AI wait for a person to approve
- Promote from development to QA to production
Why it matters: teams move fast without breaking trust in the numbers.
Talk with your data, like you talk with family.
No query language, no waiting for a report. Ask Everything in your own words, or just speak to Sarah. Your data answers back with the numbers, the reasons and what to do next, and it remembers what you were talking about.
- Ask in plain words, by chat or voice"How did we do last week?" works as well as any SQL.
- Answers you can seeTiles, charts, maps and forecasts, ready to drill into, share or schedule.
- Follow-ups and what-ifsKeep the conversation going: "why?", "show me by store", "what if prices rise 4%?"
- Only what you may seeEvery answer follows your permissions, and shows the tables it came from.
AI made for SAP ECC.#
Your SAP ECC system holds years of material movements and accounting documents. DataSimple brings them into your lakehouse and turns them into answers you can ask for in plain words.
Which materials have not moved in the last 6 months, by plant?
Slow and non-moving stock by plant and storage location, from goods movements.
What will we consume next quarter for our top materials?
A forecast of consumption with a likely range, from goods issues over time.
Which vendor invoices look like possible duplicates?
Same vendor, amount and reference within a short window, flagged for review.
Where did spend rise most this year, by cost centre and GL account?
Top changes against last year, with the line items behind each.
All the hard parts, already built.
Replication, storage, modelling, security and analytics, made for the AI era.
Every source into Delta Lake. One full copy, then only the changes.
Connect as many source systems as you need and lift and shift each one with delta loads. Deletes are flagged, late rows are caught, and every run is recorded.
Every version kept.
Silver keeps current values and full history.
Delta Lake, Iceberg, Parquet.
Choose one or all three. Any tool can read them, today and later.
Rules from a sentence.
datediff(shipped_date, promised_date)
Set once. Enforced everywhere.
Grants, row filters, column masks, quality rules and lineage apply in SQL, Python, reports and voice.
Talk to your data.
Sarah opens the reports you may see, reads out what changed, and answers follow-ups.
“Sarah, why is Europe below plan this month?”
Test a decision before you make it.
Example model: plan $40.0M and 1.22M units. In DataSimple the numbers come from your Gold tables.
One installation serves many companies or departments, as if each had its own.
Each tenant has its own users, rules and data, and data never crosses tenants. A new department or client is a new tenant ID, not a new installation.
From the first source system to next year's plan.
Book a build sessionReplication, storage, modelling, conversation, forecasting, demand planning and financial planning, all on the same governed lakehouse. Everyone sees the same numbers, and every plan is built on real data.
See what's coming, with a range you can trust.
Ask for a forecast on any Gold measure. Several methods, including a pretrained forecasting model, are tested on your own history, and the best one is shown with its likely range.
- Forecast any measure by asking
- Accuracy shown on every forecast
- Refreshes itself on a schedule
One demand plan everyone agrees on.
Start from a forecast, add what sales and marketing know, plan promotions with their expected lift, and agree one number. Every change is tracked and its value measured.
- Baseline by product and customer
- Promotion planning and approval
- Alerts when actuals move off plan
Budgets and forecasts on live data.
Build the budget, keep a rolling forecast, and test scenarios by chat or voice. Plans are saved to the lakehouse with versions and approvals, next to the actuals.
- Budget, rolling forecast, what-if
- Versions and approvals
- Plan versus actual, explained
Your cloud. Your formats. Your rules.
What DataSimple does today, and what is coming next, to keep your data safe and make it easier to build on and use.
In your own cloud
Data is stored in your cloud account, never in someone else's system.
Governed once
Who sees which rows and columns is set in one catalog and applied everywhere.
Every change traced
Lineage, audit columns on every row, and a history of every run and change.
AI with guardrails
Assistants read only governed data, and changes wait for a person to approve.
See the four checks on every request, in 3D
More secure. Simpler to build. Simpler to use.
Protection that follows the dataMore secure
PII detection and maskingAvailable
Personal data is found while profiling, tagged, and masked for anyone not allowed to see it.
Row filters and column masksAvailable
Each person sees only their rows and columns, everywhere.
Policies by tagComing soon
One rule such as "mask all PII except for HR" covers every tagged column.
Time-limited accessComing soon
Request a table, the owner approves, access ends on its own.
Access reviews and unusual-access alertsComing soon
Quarterly reviews, and alerts on large exports or odd access.
Right to be forgotten and retentionComing soon
Delete a person everywhere for GDPR; purge data by age automatically.
Your own encryption keys and an evidence packComing soon
Customer-managed keys, and one export for SOC 2 or ISO 27001 reviews.
Less setup, fewer stepsSimpler to build
Suggested quality rulesAvailable
Rules proposed from your data, accepted with one click.
Impact before you changeAvailable
See every table and report downstream before editing a table.
Pipelines from instructionsAvailable
Ask Everything to create a pipeline and schedule it.
Source auto-setupComing soon
Keys, change columns and change-capture method applied for you on connect.
"Describe it, get it built"Coming soon
One request drafts the Silver and Gold tables, rules and schedule to approve.
Industry starter kitsComing soon
A ready Manufacturing or Retail lakehouse with reports, in one click.
Answers where people already workSimpler to use
Data quality scoreAvailable
Every table scored on completeness, validity, freshness and more.
Excel export and Teams / Slack alertsAvailable
Export results to Excel; quality alerts to Teams or Slack.
"Tell me when…" alertsComing soon
Plain words become an alert: "tell me when Europe drops 5% below plan".
Reports in Teams, Slack and mobileComing soon
Everything reports and Sarah wherever your team works.
Live Excel link and data productsComing soon
Governed numbers refreshed in Excel; certified data sets others can request.
Trust badge on every answerComing soon
Quality score, freshness and source tables on every report card.
Security and compliance
Enterprise-grade security and compliance certifications
Questions, answered.
01Where is my data stored?
In your own cloud account: Amazon S3, Azure Data Lake Storage or Google Cloud Storage, never in someone else's system. It is stored as Delta Lake, Apache Iceberg or Apache Parquet, open formats any engine can read, so there is no lock-in.
02Do I need to know data models to use it?
No. Anyone can start using DataSimple without prior knowledge of data models. Pick your formats, and DataSimple builds the rest: replication with change capture, Bronze, Silver and Gold tables, security and a conversational AI you can simply talk with.
03Which source systems can it replicate?
As many as you have: SQL Server, Oracle, Db2 and COBOL systems, SAP ECC, Salesforce, APIs, files and Kafka. A backup (.bak or dump) is restored and loaded for you. The first run copies everything; after that only new, changed and deleted rows move.
04How fast is it?
In one measured run, a 150-million-row table was read from the source and written as Parquet, Delta Lake and Iceberg, at about 280,000 rows every second end to end. A 1.5-million-row table took 41 seconds, start to finish. Speed depends on the source, network and compute size.
05What does it need to run?
DataSimple runs on a small 2 vCPU, 8 GB application server. Large tables are processed in parallel by a load process that scales on demand and steps away when the load is done.
06Can the AI assistants change my data?
No assistant has privileges of its own: it acts as you and reads only governed data. Every change waits for a person to approve.
07Can several teams or companies share one installation?
Yes. One installation serves many companies or departments, as if each had its own. Each tenant has its own users, rules and data, and data never crosses tenants. A new department or client is a new tenant ID, not a new installation.
08We run SAP ECC. Where do we start?
We start from MSEG (material document items) and BSEG (accounting document line items), and we can show you sample AI results in the meantime. Before you start, get permission from SAP for extracting and using data from your ECC system under your licence.
Your first lakehouse, built this week.
Bring one source system. We connect it, replicate it, model a Gold table with your rules, and you ask your first question.