Blog
Keep the original data. Define your own answers.
Writing on how a data platform is built: ETL, ELT, data integration, data lakes, dbt and BI — drawn from Undercroft's own source code.
Integrations
Articles
MCP server for business data: Xero, Gmail and Drive
Use an MCP server for business data to work with Xero, Gmail, Drive and HubSpot through Undercroft, with personal permissions, grants and agent skills.
Data integration with REST APIs and YAML connectors
Learn how data integration works with declarative REST API connectors: YAML specs, auth, pagination, watermarks, rate limits, and visible failures.
ETL vs ELT: differences and why raw data comes first
ETL vs ELT explained with Undercroft: compare pipeline designs, choose when each fits, and see how keeping raw data lets you rebuild changing business models.
Gmail integration: from email and attachments to SQL
Build a Gmail integration that lands chosen-label emails and attachments in a raw data lake, extracts text, and makes it queryable in Postgres with SQL.
Google Drive integration: OCR and document search
Use Google Drive integration to ingest selected folders into a raw data lake, OCR Vietnamese scans, classify documents and search their extracted text.
HubSpot integration: from CRM to Postgres and dbt
Build a HubSpot integration that syncs CRM objects and custom properties into a raw lake and Postgres, then joins HubSpot with Xero in dbt.
Raw data lake: immutable storage on S3 and MinIO
Learn how an immutable raw data lake on S3 or MinIO preserves source bytes, deduplicates content, and lets you rebuild Postgres tables and dbt models.
Undercroft: an open-source data platform for self-hosted ELT
Explore Undercroft, an MIT-licensed open-source data platform combining ingestion, a raw lake, dbt and BI. Compare its scope with Fivetran and Airbyte.
Xero integration: sync accounting data into Postgres
Build a Xero integration with OAuth, YAML connectors, an immutable raw lake and Postgres. Learn how incremental sync, full re-sync and dbt fit together.
Xero reporting with SQL and dbt: build your own views
Build custom Xero reporting with SQL and dbt: P&L, receivables ageing and cash views, using exact decimals and keeping missing amounts visibly missing.