dbt Labs licensing covers the terms under which dbt Labs, LLC makes dbt available. dbt is a framework that transforms data inside a warehouse with SQL models, tests and documentation. It reaches users through four routes: an Apache 2.0 open-source package, a free proprietary binary, a hosted subscription product and usage-based add-ons. dbt Labs became a wholly owned subsidiary of Fivetran when the two companies completed an all-stock merger on 1 June 2026.[4][5]
The four routes
| Route | Licence or contract | Cost | Deeper article |
|---|---|---|---|
| dbt Core v1.x | Apache 2.0 | Free | Open source and product licensing |
| dbt OSS (open-source subset of dbt v2) | Apache 2.0 | Free | Open source and product licensing |
| dbt (full v2 distribution) and the VS Code extension | dbt Product Licensing Agreement | Free to install and run | Open source and product licensing |
| dbt platform | Terms of Service, Order Form | Plans from free to custom | Plans, seats and usage |
| dbt State, dbt Wizard, dbt Copilot | Service Consumption Table, Terms of Service | Usage-based | State, Wizard and contract terms |
The documentation states that “dbt v1 is released under the Apache 2.0 license” and “dbt v2 is proprietary to dbt Labs, made available under the dbt Product Licensing Agreement.” Both are free, and dbt v2 is “still completely free to use, with some features unlocked by a free login or a paid dbt platform account.”[1]
What changed in 2026
dbt Labs’ licensing FAQ records the shifts. In May 2025 dbt Labs announced the next-generation Fusion engine, whose proprietary distribution then used the Elastic License 2.0. On 1 June 2026 the company removed ELv2 references because the Fusion repository moved into the dbt-core repository and was relicensed to Apache 2.0, introduced the dbt Product Licensing Agreement for the Fusion binary, and noted that MetricFlow had moved from the Business Source License to Apache 2.0 in October 2025.[2] On 15 September 2026, as dbt v2 reached general availability, the FAQ renamed the Fusion engine as dbt and dbt Core v2 as dbt OSS and declared that the dbt Core name no longer applies to any v2 distribution.[2] The merger announcement of 1 June 2026 described the same open sourcing as the release of dbt Core v2.0 under Apache 2.0.[4]
An asset manager should therefore treat older references, such as “ELv2 Fusion engine” or “dbt Core 2.0 beta”, as historical.
Plans and metrics
The dbt platform has four plans. Developer is free with one developer seat, 3,000 successful models built per month and one project. Starter costs USD 100 per user per month for five developer seats, 15,000 models, 5,000 queried metrics and one project. Enterprise and Enterprise+ have custom pricing, with 100,000 models and 20,000 queried metrics, 30 projects on Enterprise and unlimited projects on Enterprise+.[3] Usage is measured as Successful Models Built, plus Semantic Layer Queried Metrics, while seats are billed mainly on Developer and Read licences.[8]
dbt State, which reuses models that do not need rebuilding, and dbt Wizard, an AI agent for analytics engineering, are add-ons with their own metrics: daily active target tables and tokens converted into dollars.[7]
Contract documents
A dbt platform customer is bound by the Terms of Service updated on 16 October 2025 and an Order Form, with Order Terms of twelve months unless otherwise agreed.[6] The Service Consumption Table (effective 1 July 2026) applies when a customer uses consumption add-ons.[7] Customers who buy dbt Labs services under a Fivetran Master Subscription Agreement are covered by Fivetran’s Requirements for dbt Labs Services, which replace the Fivetran consumption pricing provisions and support policy for those services.[5] The free binary is governed by the dbt Product Licensing Agreement, which a user accepts by first downloading or using it.[9]
Terminology map
dbt Labs’ names changed between 2025 and 2026, so older inventories and contracts use different words for the same software. The FAQ and documentation give this mapping.[1][2]
| Older name | Current name | Licence |
|---|---|---|
| dbt Core 1.x | dbt Core v1.x (unchanged) | Apache 2.0 |
| dbt Core v2 (beta, open source) | dbt OSS | Apache 2.0 |
| dbt Fusion engine | dbt (full distribution) | dbt Product Licensing Agreement |
| dbt Cloud, dbt Cloud Terms of Use | dbt platform, Terms of Service | Order Form and Terms of Service |
| State-aware orchestration | dbt State | Usage-based add-on |
| dbt Starter, previously Team | Starter plan; Team before 2023-08-11 is the legacy plan | Order Form or card terms |
The FAQ says the dbt Core name “no longer applies to any dbt v2 distribution”, and that the rename is a rebrand of the same public code for existing beta users.[2] Customers should expect procurement records to mix old and new names for some time.
How dbt relates to Fivetran products
Fivetran sells its own Transformations product, metered in monthly model runs under Fivetran’s consumption table; see Fivetran Transformations, Activations and AI features. The Fivetran requirements for dbt Labs Services say that Fivetran’s consumption-based pricing provisions do not apply to dbt Labs Services, which are paid at the Order Form fees. A customer that uses both a dbt platform subscription and Fivetran Transformations therefore holds two separate metering relationships even though the companies merged.[5]
Points for software asset managers
- Inventory which distribution each developer machine and CI runner uses: dbt Core v1.x, dbt OSS, or the proprietary dbt binary with its Product Licensing Agreement and telemetry.
- Hosted or embedded use of the proprietary dbt for third parties brings the redistribution conditions; internal use and consultant-run standalone deployments do not.[2]
- Seat counts, model builds and add-on usage on the dbt platform follow different rules per plan; Enterprise tiers draw on an annual commitment.[8]
- Contracts may combine Fivetran and dbt Labs paper after the merger; confirm which agreement governs each service.
Out of scope
Partner and reseller terms, dbt Labs professional services and training statements of work, third-party model provider terms for dbt Wizard, and the warehouse costs incurred when dbt builds models are not covered.