dbt Cloud vs dbt-core: Which One Should You Use in 2026?

If you're building data pipelines with dbt, you've probably asked yourself: should I use dbt Cloud or stick with dbt-core?

It's a fair question. Both are built by dbt Labs, both run the same dbt models, and both can get your transformations into production. But they solve different problems, and the right choice depends entirely on what you actually need.

What is dbt-core?

dbt-core is the open-source command-line tool. You install it locally (or in a container), point it at your data warehouse, and run dbt build. It handles the SQL compilation, dependency resolution, testing, and documentation.

What it doesn't do: schedule itself, manage credentials securely, provide a web UI, or run in production without additional infrastructure.

What is dbt Cloud?

dbt Cloud is dbt Labs' commercial SaaS platform. It wraps dbt-core with a browser-based IDE, job scheduling, CI/CD integration, a semantic layer, and team collaboration features. It's a full development environment.

The Real Comparison

Let's cut through the marketing and look at what actually matters.

Scheduling & Production Runs

dbt Cloud includes built-in job scheduling. Define a job, set a cron, done. It also handles CI runs on pull requests.

dbt-core has no built-in scheduling. You need to bring your own orchestrator — Airflow, Prefect, Dagster, GitHub Actions, a cron job, or a managed service like ModelDock.

Development Experience

dbt Cloud provides a browser-based IDE with syntax highlighting, auto-complete, lineage visualization, and the ability to run models directly from the browser.

dbt-core users typically work in VS Code (with the dbt Power User extension), their terminal, or any editor they prefer. Many teams prefer this because they already have established local development workflows.

Pricing

This is where it gets interesting.

dbt Cloud pricing starts free for one developer seat, but scales quickly. The Team plan is $100/seat/month. For a team of 5, that's $6,000/year — and that's before you hit the Enterprise tier for features like SSO or audit logging.

dbt-core is free forever. But "free" is misleading — you're paying in engineering time. Setting up Airflow, writing Dockerfiles, managing credentials, building CI/CD pipelines, and maintaining all of it has a real cost.

The middle ground: Managed dbt-core schedulers like ModelDock start at $0/month (free tier) and top out at $249.99/month for unlimited everything. You get production scheduling without the infrastructure burden or the dbt Cloud price tag.

Adapters & Warehouses

Both support all major warehouses: Snowflake, BigQuery, PostgreSQL, Redshift, and Databricks.

dbt-core additionally supports community adapters for warehouses that dbt Cloud doesn't, including Microsoft Fabric, DuckDB, and dozens of others.

Credential Security

dbt Cloud manages credentials within their platform. Your warehouse passwords and tokens are stored on dbt Labs' infrastructure.

dbt-core keeps credentials wherever you put them — typically profiles.yml or environment variables. With a managed scheduler, credentials can be encrypted at rest (ModelDock uses AES-256-GCM) and only decrypted inside isolated containers at run time.

When to Choose dbt Cloud

dbt Cloud makes sense when:

• Your team needs a browser-based IDE and doesn't want to set up local development
• You want CI/CD on pull requests out of the box
• You need the semantic layer for BI tool integration
• Budget isn't the primary constraint
• You want a single vendor for development and production

When to Choose dbt-core

dbt-core makes sense when:

• You already have a local development workflow you're happy with
• You want to use community adapters (Fabric, DuckDB, etc.)
• You want full control over your infrastructure and credentials
• You need to keep costs low, especially as the team grows
• You prefer open-source tools

The Scheduling Gap

The biggest challenge with dbt-core isn't dbt itself — it's everything around it. Your models work great locally, but getting them running on a schedule in production means:

1. Setting up an orchestrator (Airflow, Prefect, etc.)
2. Writing Dockerfiles for your dbt version + adapter
3. Managing profiles.yml and credentials securely
4. Building CI/CD pipelines for deployment
5. Setting up monitoring and alerting
6. Maintaining all of the above indefinitely

This is where managed dbt-core schedulers fill the gap. They handle the infrastructure so you can focus on writing models.

Our Take

If you need a full development environment and the budget supports it, dbt Cloud is excellent. If you just need your dbt-core project running on a schedule — reliably, securely, and without managing infrastructure — there's a simpler path.

ModelDock was built specifically for this: connect your repo, pick your adapter, set a schedule. Your dbt-core project runs in production. No Airflow to manage, no Docker to wrangle, no CI/CD to build.

Free during open beta. No credit card required.