The 3 Best Coalesce Alternatives in 2026

dbt is where the T in ELT happens: you write SELECT statements, dbt handles dependency order, materialisation, testing and documentation, all in Git. dbt Core is Apache-2.0 and free to run from any scheduler you already have; dbt Cloud adds a browser IDE, job scheduling, the semantic layer and a catalogue, priced per developer seat.

Best for: analytics engineering, SQL models, testing

• The de-facto standard for warehouse transforms
• dbt Core is free and self-hosted
• Tests and lineage come with the model

SQLMesh plans and runs SQL (and Python) models like dbt does, but it understands the SQL it runs. That lets it create isolated development environments without copying tables, skip rebuilding models that haven’t changed, run unit tests without touching the warehouse, and catch errors before execution across 10+ SQL dialects. It reads existing dbt projects. SQLMesh is Apache 2.0 licensed and the README describes it as a Linux Foundation project; there is no paid tier listed in the repository.

Best for: virtual environments, open source, dbt-compatible

• Dev environments without warehouse copies
• Never rebuilds an unchanged table
• Compile-time SQL checks

Dataform lets you define BigQuery tables, views and assertions in SQLX files kept in Git, and runs them in dependency order. Google lists Dataform as a free service: the charges come from BigQuery for the queries it executes and from Cloud Logging, which is required for workflow invocations. Scheduling can add Cloud Composer, Cloud Scheduler or Workflows costs. It only targets BigQuery.

Best for: BigQuery, Google Cloud, no licence fee

• No licence fee at all
• Lives in the BigQuery console
• Git-backed SQLX projects