xREF DataMapper is an AI-powered source-to-target mapping engine that automates schema discovery, intelligent field mapping, confidence scoring, and change governance — so you can migrate thousands of data assets before writing ETL.
From schema discovery to audit-ready exports — one platform, end to end.
Multi-agent pipeline with Claude, GPT, DeepSeek & custom LLMs. Exact, fuzzy, and semantic matching with confidence scores per field.
Get a readiness verdict per column, table, and program before writing a single line of ETL. Catch type mismatches, missing keys, and gaps early.
Role-based access, human approval gates, versioned mappings, and full audit trails. Every action is logged — who, what, when, from where.
Export mapping specs as CSV, Excel, HTML reports, and ready-to-run SQL DDL for BigQuery, Snowflake, Redshift, and more.
A 4-stage agentic pipeline takes your source schemas and produces production-ready target mappings.
Discover and profile source schemas from CSV, DDL, Excel, or live databases. Infer data types, keys, and relationships.
Parse your target schema (BigQuery, Snowflake, etc.) and build a canonical metadata model for matching.
AI agents match every source field to the target using exact, fuzzy, and semantic strategies. Confidence scored per mapping.
Generate the mapping spec, readiness report, lineage graph, and dialect-specific SQL — ready for review and deployment.
Security, compliance, and governance are first-class — not afterthoughts.
All secrets, credentials, and connection strings are encrypted with Fernet (AES-128-CBC). Keys never touch disk unencrypted.
Every workspace is a fully isolated tenant. Data, sessions, and mappings are never shared across tenants.
AI suggests mappings; humans approve. Low-confidence mappings cannot be auto-approved. Full traceability on every decision.
Support for OpenID Connect single sign-on. JWT-based sessions with httpOnly cookies, CSRF protection, and forced HTTPS in production.
Start migrating your data assets with AI-powered precision. Create a workspace and run your first mapping in minutes.
Get Started →