The platform by Data Reply

Metabasis

The AI data development lifecycle in one workspace: discover, assess, design, build, validate and operate, with agents doing the work and your team approving it.

Sign inCreate an account
Or book a 30-minute walkthrough

What is inside

Twelve working areas, grouped by the stage of the lifecycle they serve.

1. Discover

  • Ingestion and discovery

    Read-only connectors, schema discovery, files and buckets

  • Datasets and profiling

    Column profiles and the issues they reveal

2. Assess

  • Readiness assessments

    Scored, evidenced, with ROI per use case

  • BI migration inventory

    Every Qlik and MicroStrategy asset triaged

3. Design

  • Modelling

    Dimensional, Data Vault and gold-layer models with DDL

  • Mapping and contracts

    Source-to-target mappings and data contracts

4. Build

  • Pipeline

    Code per medallion layer for your target platform

  • Conversion

    Legacy SQL, SAS and .NET to the target dialect

5. Validate

  • Data quality

    Rules, validation runs and run comparison

  • Testing and reconciliation

    Migrated numbers checked against the source

6. Operate

  • Lineage and observability

    From code and BI scripts, with run health

  • Governance and audit

    Roles, policies and every change recorded

See it working

Real screens, on a synthetic 5,000-app Qlik estate and a sample project.

The 4-minute demo: profiling, quality rules, validation and modelling on a sample project.

Metabasis · Assess
Metabasis BI migration inventory: a Qlik app marked complex, with its complexity drivers, ten data sources and its Amazon Quick object mapping
Every app in the estate triaged, with the reasons: what drives the effort, which sources it reads, and how each visual maps to Amazon Quick.

Architecture

Deterministic code for the facts, reviewable agents for the judgement, and outputs in open formats for the platforms you already run.

The full architecture
Metabasis architecture: inputs flow into the engine, which produces outputs for your warehouse and Amazon QuickWhat you bringLive databases (read-only)Files and cloud bucketsQlik and MicroStrategy exportsSQL, SAS and .NET codePolicies and strategy documentsMetabasisDeterministic coreParsers, profilers, lineage, scoringReviewable agentsModels, rules, conversion, plansQueue workersResumable, chunked, estate-scaleGoverned metadataRoles, row-level security, auditWhat you getModels and DDLPipeline code per layerQuality rules and contractsLineage and documentationAmazon Quick starter packsEvidence packs and reportsWhere it runsRedshiftSnowflakeDatabricksSynapseAmazon Quick

Works with

Reads from

  • PostgreSQL
  • MySQL
  • SQL ServerSQL Server
  • Snowflake
  • Databricks
  • Google BigQueryGoogle BigQuery
  • Amazon S3Amazon S3
  • Azure Blob StorageAzure Blob Storage
  • Google Cloud StorageGoogle Cloud Storage

Migrates BI from

  • Qlik
  • MicroStrategy

Builds and delivers to

  • Snowflake
  • Databricks
  • Amazon RedshiftAmazon Redshift
  • Azure SynapseAzure Synapse
  • Amazon QuickAmazon Quick

Security and access

Hosting and data residency are agreed per engagement.

Roles per project

Owner, admin, member and viewer, plus a client role limited to answering assessments.

Read-only by default

Connectors only read. Uploads go to project-scoped storage behind signed links.

Your choice of model

Claude by default; a project can use its own endpoint, such as Azure OpenAI.

Audit log

Decisions and changes are logged per project for owners and admins.

Ready when you are

Sign in, create an account, or let us show you Metabasis on your own problem.

Create an accountBook 30 minutes
The platform | Metabasis