Virtualize
Test databases without copy tickets
Create isolated VDBs from production snapshots without making every team wait for a full database copy.
[ SOURCE ][ MASK ][ VDB ][ REFRESH ]
SOFI Test Data for private databases
Realistic test data without exposing production.
SOFI runs inside your environment to mask, virtualize, and refresh databases for dev, QA, demos, and analytics. Teams get production-like data; sensitive values stay under your control.
or start free on SaaS — no card · 1,000 free credits
< 60s
provision
95%
less storage
0
raw data leaves
private test data run
sofi engine
$
sourcequeued
scanqueued
riskqueued
savedqueued
Detected fields
runningtables
42
sensitive
17
customer.emailPIIhigh
cards.numberPCIhigh
orders.totalsafekept
tenant.idjoin keypreserve
Raw production values stay inside your environment.
Built for private data teams
Runs inside your environmentProduction data stays privateRead-only source accessConsistent masking rulesAudit trail for every refresh
connectors for the databases you already run
Connect production sources once. Reuse them for every safe test environment.
01 / product surface
The test data workflow your teams keep rebuilding by hand.
SOFI combines source connectivity, masking, virtualization, refresh, and audit controls in one on-prem platform for lower environments.
Mask
Mask sensitive data before teams use it
Replace PII with realistic values while preserving formats, joins, constraints, and application behavior.
Refresh
Keep lower environments current
Refresh dev, QA, demo, and analytics databases without repeating slow export, copy, and sanitize jobs.
Operate
Operate under your security controls
Deploy in your VPC, private cloud, or bare metal network with SSO, RBAC, audit logs, and private routing.
02 / discover · mask · comply
See the masking flow in motion.
Sensitive columns are discovered, replaced with realistic values, and signed off against the compliance frameworks your auditors care about.
Sensitive Discovery
Scanning…
FULL_NAMEMatheus Vilarino
EMAIL_ADDRmath@sofi.io
CPF_DOC522.345.678-09
PHONE_NUM+55 11 98888-7777
CREDIT_CARD4532...1122
ZIP_CODE04533-010
SOFI Masking Hub
Compliant Data
Waiting…
FULL_NAME
EMAIL_ADDR
CPF_DOC
PHONE_NUM
CREDIT_CARD
ZIP_CODE
Certified Standards & Regulations
GDPR
LGPD
HIPAA
PCI
SOC2
03 / developer first
A CLI and API surface your platform team can standardize.
SOFI can be driven from the dashboard, CLI, REST API, CI/CD, or agent tooling. The important part is that every path enforces the same masking, RBAC, TTL, and audit controls.
CLI
REST
Webhooks
$ sofi workspace create
04 / workflow
From production source to safe test database.
SOFI connects to the source, detects sensitive fields, applies masking, and provisions a usable database for the team that requested it.
01source ready
Connect the production source
Register the database engine, route, and read-only credentials inside the customer network.
02risk map
Find sensitive columns
Profile schemas for PII, relationships, tenant keys, formats, and fields that need protection.
03policy active
Apply masking rules
Use deterministic rules so masked data keeps joins, constraints, dates, totals, and formats intact.
04VDB ready
Create the test database
Provision isolated VDBs for dev, QA, demos, analytics, or pull-request environments.
05refresh active
Refresh safely over time
Keep test environments current without reopening tickets or exposing raw production values.
05 / real-time data flow
Continually sync. Continually save.
Log-based CDC keeps VDBs in step with production. CoW thin clones turn a 10TB source into a few TB of unique blocks — zero impact, maximum consistency.
Source Cluster
StatusProduction Live
Physical Volume
10.0 TB
Block usageCritical (85%)
Data Timeflow
Space Saving10.0 TB
Physical: 10TBVirtual: 10.0TB
06 / deployment
Built for teams that cannot send production data to a SaaS.
The control plane, workers, masking jobs, snapshots, and VDB provisioning run inside your VPC, private cloud, or bare metal environment.
SSO + RBAC
Private registry
Audit exports
Storage policies
07 / integrations
Use the tools already in your delivery path.
SOFI integrates with the systems that request, approve, run, observe, and revoke lower-environment data.
connected
CI/CD
Create masked workspaces for pull requests, release branches, QA windows, and demos.
GitHub ActionsGitLab CIJenkins
connected
Identity
Use SAML, OIDC, LDAP, and RBAC so data access follows enterprise controls.
SAMLOIDCLDAP
connected
Automation
Send provisioning, refresh, policy, and audit events to internal tools through webhooks.
RESTWebhooksAPI keys
connected
AI agents
Expose safe test-data operations as typed tools while keeping masking and audit enforced.
MCPOpenAICursor
08 / use cases
Where production-like data is worth the controls.
Teams adopt SOFI where fake fixtures are too weak, raw production is too risky, and manual database copies are too slow.
Pull-request databases
Give each review app a short-lived masked database that behaves like production.
QA refresh cycles
Refresh lower environments from approved snapshots without copy tickets.
Audit-ready sandboxes
Let analytics and support teams work with realistic data under policy evidence.
Platform self-service
Move database requests from manual scripts into governed CLI, API, and dashboard flows.
// 06 / agent ready //
Built for AI agents too.
Expose safe test-data operations as typed tools to Cursor, Claude Code, Windsurf, or any MCP-compatible agent. Masking, RBAC, audit — all still enforced.
MCP Server
Skill file
Typed tools
RBAC inherited
~/.cursor/mcp.json
{
"mcpServers": {
"sofi": {
"command": "npx",
"args": ["@sofi/mcp", "--workspace", "acme"],
"env": { "SOFI_API_KEY": "${SOFI_API_KEY}" }
}
}
}cursor · agent prompt
liveuser> Spin up a masked dev DB for PR #482
agent> Calling sofi.workspaces.create with source=prod_postgres, policy=checkout_pii…
✓ ready at postgresql://checkout-pr-482.sofi.local:5432/app
09 / operating model
Stop waiting days for a safe database copy.
Replace manual refresh tickets, full copies, and after-the-fact sanitization with repeatable masked VDBs for every team.
test data refreshsafe
source snapshotcaptured
masking policyapplied
test databaseready
Manual DB copy01
4 days
tickets, scripts, approvals
Mask after copy02
18 min
sensitive data exists before cleanup
SOFI masked VDB03
< 60 sec
virtualize, mask, refresh
[ MASK ][ VIRTUALIZE ][ REFRESH ][ AUDIT ][ COMPLY ]
// ready for safer test data
Give teams useful test databases without handing them raw production data.
no card on SaaS · 14-day on-prem pilot