Transformations in Weld

Weld supports two approaches to transformations:

  1. Weld Transform: build and schedule models directly in Weld with the SQL editor.
  2. Third-party orchestration: plug in dbt/dbt Core or your warehouse/BI tools while Weld handles ingestion, orchestration, and reverse-ETL.

Weld Transform

Weld’s SQL editor connects directly to your destination (warehouse), so you can write, test, and run SQL models inside Weld.
Perfect if you want one place to build, schedule, and collaborate.

ℹ️ Read the primer on Data Models: drafts, published models, table references, and more.

Key capabilities

  • AI assistant: get SQL help and debugging from Ed
  • Data Lineage: visualize dependencies and impact
  • Version history: track, diff, and roll back changes
  • Scheduling: run models automatically
  • Realtime collaboration: build together with presence and safe concurrency

Guides

Third-party orchestration

Prefer external tools? Weld integrates them into your workflows so you can keep your stack and still get centralized orchestration.

Why choose this path

  • Use familiar tools (dbt, dbt Core, Power BI, Fabric, Looker Studio, Snowflake/BigQuery SQL)
  • Weld handles ingestion and reverse-ETL, plus consistent scheduling and monitoring
  • Fewer handoffs and full visibility across the pipeline

Integrations

  • dbt Cloud: trigger jobs from Weld with webhook-based orchestration.
    Read setup →

  • dbt Core: run Core externally while Weld powers ingestion and activation.
    Read setup →

  • Other / Webhooks: keep transformations in your warehouse or BI tools.
    Read setup →

✅ Already invested in dbt/BI/SQL? Weld doesn't replace those tools. It connects and orchestrates them within your existing workflow.

See the full list here: All Destinations. Common pairings we see:

  • Snowflake: often with dbt/dbt Core. BI via Power BI or SQL-native dashboards.
  • BigQuery: often with dbt/dbt Core. BI via Looker Studio or Power BI.
  • Databricks: often with dbt/dbt Core. BI via Power BI/Fabric.
  • Postgres: lightweight analytics via SQL-based BI or embedded dashboards. Orchestration with dbt Core.
  • Amazon S3 (lake/landing): transformed into your warehouse or queried via warehouse engines. Orchestrated alongside dbt/dbt Core.

(Check the Destinations page for everything Weld supports.)

Quick comparison

Use when…Weld TransformThird-party orchestration
You want one place to model & schedule✅—
You already use dbt/BI for transforms—✅
You need centralized lineage/versioning in Weld✅Partial (depends on tool)
You want to keep vendor-native modeling (e.g., Fabric/Power BI)—✅

Transforms API

Need to manage models programmatically? The Transforms API lets you create, update, publish, and version data models through HTTP requests.

EndpointWhat it does
GET /transformsList models with filters for status, name, and folder
POST /transformsCreate a new model with SQL and optional auto-publish
PATCH /transforms/{id}Update SQL, name, folder, or materialization
DELETE /transforms/{id}Soft-delete a model and remove its DWH table/view
POST /transforms/{id}/publishMaterialize the model to your warehouse
GET /transforms/{id}/versionsBrowse the SQL revision history
POST /transforms/bulkCreate up to 50 models in one request

Common use cases:

  • CI/CD deployments: push SQL models from Git to Weld as part of your release pipeline
  • Bulk migration: import models from another tool using the bulk endpoint
  • AI agents: let coding assistants (Claude Code, Cursor, Copilot) create and publish models via the API
  • Notebooks: manage models from Jupyter or Python scripts

The Transforms API is available on the Business plan.

Read the full Transforms API blog post →

Explore the API reference →

Import & Export

Prefer working with files? Weld has built-in Model Import & Export for moving models between workspaces or converting to/from dbt format. This complements the API for offline workflows.

Next steps

Was this page helpful?