DataForge discovers sensitive data with AI, masks it on the way out, slices it down to the size your team needs, and spins up safe, on-demand environments for dev, test, and analytics. Engineers stop waiting on data. Compliance stops worrying about it.
Four sentences that come up on every release cycle — and the answers your team is forced to give without proper tooling for the data itself.
We can't give devs production data.
Compliance won't allow it. Legal won't allow it. But synthetic data doesn't reproduce the bugs, and a half-empty test database doesn't catch the real ones. Development happens against data that doesn't reflect reality — and bugs ship to production anyway.
Refreshing the test environment takes a week.
The DBA pulls a copy, scrubs it, loads it, verifies it. By the time the test environment is fresh, the data is already two sprints stale. Multiply by every team that needs a copy.
We don't actually know where our PII lives.
There's PII in the database somewhere — there always is. You can't list every column that holds it. Neither can your auditor. That's a problem before it's a problem.
Test data was an Excel file the team passed around.
Each environment looks different. Each developer's local copy looks different again. Bugs that reproduce in staging vanish in dev. Nobody trusts what they see.
Four operations that today live in scripts, tickets, and DBA brains — built into one product. The full lifecycle of non-production data, from "where is the PII" to "the new dev has a working environment."
DataForge scans your schemas and flags PII, payment data, national IDs, and patterns you define. Smarter than pure regex; faster than a manual audit. The AI's reasoning is visible and reviewable — run the model on your own hardware for full sovereignty, or plug in your LLM provider of choice (Gemini, Claude, OpenAI) when the data path suits your policy.
Choose the runtime: a local model on your hardware for air-gapped sovereignty, or your preferred cloud LLM when the data path is acceptable to your privacy office.
Every finding has a confidence score and a reason. You approve, reject, or refine — nothing is masked until you say so.
Replace sensitive values with realistic ones that preserve format, length, and referential integrity. Foreign keys still match. Reports still run. Bugs still reproduce. PII no longer exposed.
Pull subsets by tenant, customer, date range, or any rule that fits your workload. Smaller environments, faster refreshes, lower storage costs — and the relationships between tables stay intact.
Provision a fresh, masked, right-sized database environment for a developer, a tester, or a CI job. Use it for as long as you need. Expire it on a schedule or tear it down on command. No more shared dev databases that nobody is brave enough to refresh.
Personal namespaces for engineers. Reproducible slices for QA. Fresh, governed copies for every CI run — without the DBA being the bottleneck.
Create on demand or on a schedule. Refresh, expire, recreate. The whole loop is orchestrated — no spreadsheets, no Slack reminders.
Schedule discoveries. Schedule refreshes. Trigger on commit, on ticket, on deploy. DataForge plugs into your CI/CD — it doesn't ask your team to build a new pipeline from scratch.
If your week revolves around the question "can the team have a fresh copy of production?" — DataForge is for you.
Tired of being the bottleneck when every team needs a refreshed copy. DataForge automates the work that fills your inbox.
Need repeatable, governed data pipelines into non-prod — masked the same way every time, regardless of who's asking.
Want realistic data without a week-long ticket. Realistic enough to reproduce bugs. Safe enough that nobody panics.
Need to prove sensitive data is protected outside production — with evidence, not assurances. Every mask is logged and attributable.
Answering audit questions about where regulated data lives — and producing the inventory the auditor asked for, on demand.
Three scenarios we hear on every call. Same problem, same time pressure, same audit risk — different outcomes once DataForge is in the loop.
A new developer joins the team.
Day one, they need a working environment. DataForge provisions a masked, subsetted copy of production in their isolated namespace. They're coding by lunchtime. The environment expires at the end of their project.
QA wants to reproduce a customer bug.
The bug only shows up against a specific data shape. DataForge clones the relevant production slice, masks it, and hands QA a fresh environment. The repro happens this afternoon — not next week.
An auditor asks where payment data lives.
DataForge already discovered every column. You hand the auditor the inventory and the masking rules — with confidence scores and AI reasoning. The audit takes hours instead of weeks.
Other vendors in this space want your column names and your sample data in their cloud, full stop. Avesys gives you the choice. Run the discovery model on your own hardware for air-gapped sovereignty, or plug in your existing LLM provider when the data path suits your privacy office.
Built by a team that already operates SQL Server in production through Avesys DPM. We understand the constraints of large databases, referential integrity, and operational windows — because we've worked inside them.
Run the discovery model on your own hardware for full sovereignty, or plug in Gemini, Claude, OpenAI, or your provider of choice. The architecture supports both — the choice is yours, not the vendor's.
Production data, masked data, ephemeral environments — none of it routes through an Avesys SaaS in the middle. Critical when the data being processed is the data you can't expose.
Foreign keys, system-versioned tables, columnstore, operational windows. The product was built by DBAs who've subsetted billion-row tables, not a UI team guessing at the workflow.
One vendor across monitoring (DPM), comparison (Database Compare), version control (SQL Version Control), and data management (DataForge). One operational model. One team to call.
Retail, manufacturing, finance, healthcare — environments where GDPR/KVKK/PCI obligations make casual data copying a career-ending move. Designed for buyers who can't just send their data to a cloud SaaS.
From "we'd like to give this team a copy" to "they have a fresh, masked, governed environment" — in a workflow that doesn't depend on the DBA's calendar.
Point DataForge at your source databases. Read-only by default. No agents on the SQL Server, no production write paths.
AI-assisted scanning identifies sensitive data across your schemas — with confidence scores and reasoning. You confirm the inventory.
Masking strategies, subset rules, lifecycle policies — defined once, reusable across environments and teams.
Spin up masked, subsetted environments on demand or on a schedule. Connection strings handed to whoever asked.
Create, refresh, expire, recreate. The full lifecycle is automated — and every action is logged and attributable.
One number per terabyte of data you process. No per-instance fees. No add-on modules.
€2,900 per TB, per year. Volume pricing available for estates above 10 TB. Existing Avesys customers see bundle options across the platform — talk to our team.
The questions procurement, security, and the DBA team usually arrive with on the first call. Same answers we give them.
€2,900 per TB, per year. Volume pricing available for estates above 10 TB. Existing Avesys customers can bundle DataForge into one license relationship.
Yes — DataForge is live and shipping. Pricing, support, and the full feature set are documented above. Contact us to schedule a demo against your environment.
Not in the classic sense. DataForge is purpose-built for getting safe, useful copies of production data into non-production environments — discovery, masking, subsetting, provisioning. If you need a general-purpose ETL/ELT, this isn't that.
Wherever your privacy office tells us to. The discovery model can run on your own hardware for full sovereignty, or you can plug in a managed LLM provider (Gemini, Claude, OpenAI). The runtime is configurable per deployment.
Yes — when you deploy the local AI model. DataForge follows the same on-premise design as the rest of the Avesys platform. Air-gapped installations are supported.
SQL Server is the launch target. Other engines are being evaluated based on customer demand — tell us what you run when we talk.
AI-assisted scanning combined with pattern matching, column metadata, and rules you define. The output is reviewable — every finding has a confidence score and a reason. You confirm before masking is applied.
Yes. Foreign keys still match, substitutions are consistent across tables, and format-preserving masking is the default. Reports still run, bugs still reproduce.
They're complementary. DPM monitors live SQL Server. Database Compare diffs environments. SQL Version Control versions schema. DataForge handles the data that flows between environments.
Talk to us about your estate. We'll schedule a demo against your environment with a real engineer — not a sales rep.