Credible vs. dbt

dbt is the transformation standard, and Credible works well on top of the tables it builds. It also covers that ground in the other order: write down what data means first, and Malloy is one language for transformation, modeling and materialization — so the pipeline lives inside the data model instead of underneath it.

Already decided? Migrate from dbt Semantic Layer walks through what carries over.

How dbt describes itself

What dbt is

dbt is the transformation standard, positioned after the Fivetran merger as data infrastructure for trusted AI agents: MetricFlow re-open-sourced under Apache 2.0 as a reference implementation for Apache Ossie, an MCP server shipping, and the Fusion engine in dbt Core v2.0. It increasingly runs inside the warehouse too: dbt Projects on Snowflake is generally available, and a dbt package manages Snowflake semantic views as models in your own repository.

Category

Credible: The AI Analytics Engine — you write down what data means, and it generates the stack under it
dbt: Transformation framework with an add-on metrics layer

Who it is built for

Credible: The whole organization, not one central team — ops, finance and product; spreadsheet users through ML engineers
dbt: Analytics engineers who own the transformation layer

Primary interface

Credible: The agent you already use, over MCP — plus Credible Workspaces, dashboards, notebooks and reports for the people who want a UI. Every one of them reads the same model, and the model lives in your repository
dbt: Code — models authored in SQL, Jinja and YAML through an IDE or CLI, with BI tools on top

Modeling language

Credible: Malloy — a modern programming language for data: imports, inheritance, public and private members, and queries that compose into new sources
dbt: SQL with Jinja templating plus YAML, and MetricFlow for metrics

What is open

Credible: Open core. Malloy is open source, and we build and maintain Malloy Publisher — the open-source server for Malloy models, and where our agent skills and MCP retrieval tools are open source too, runnable in any agent with no Credible account. Your model is code in your repository, and Credible is in the Apache Ossie ecosystem for semantic interchange
dbt: Open core — dbt Core and MetricFlow are Apache 2.0, the platform is commercial, and MetricFlow is a reference implementation for Apache Ossie

How the model gets built

Credible: Model first: you write down what data means, and the engine derives the pipelines, storage and serving from it. Open-source MCP tools and agent skills help you capture that meaning from where it already lives, running in your coding agent
dbt: Hand-authored models, with agent-assisted authoring

Governance

Credible: One annotation on the source, in the same file as the logic it governs — no grant objects to declare, no rule to re-attach on every surface that exposes the data, no attribute table kept in step by hand. A row filter or an authorize gate over attributes the server resolves from verified identity, and everything reading that source inherits it: dashboards, notebooks and agents alike
dbt: Lineage, tests and metric governance within the project

Materialization and caching

Credible: One annotation on the source, in the same file as the logic it materializes — no derived-table block, no rollup definitions, no refresh triggers, no orchestration run to schedule. Optional per source: query your own warehouse directly, or hold a source hot in Credible’s own storage, which is how you get fast serving without buying or banging your head against a warehouse to get it
dbt: Materializations are a core concept for transformation, but the Semantic Layer itself ships no pre-aggregation cache and compiles fresh SQL per query

How an agent finds the right data

Credible: Search. Typed targets — source, dimension, measure, view, even a dimensional value — are matched against an index of the model and come back ranked, so an agent asks for what it needs instead of picking a model and touring it
dbt: List then drill — list_metrics or get_all_models, then query_metrics or get_node_details

Where the model can be used

Credible: One model served to every surface — agents over MCP, plus APIs, an SDK, embedded dashboards, notebooks and HTML data apps. Results carry an interactive UI resource (MCP Apps, the official extension), so a client that supports it renders a real chart or table instead of the model re-narrating rows
dbt: Semantic Layer APIs into BI tools, plus an MCP server

Scale and portability

Credible: Globally distributed, high-availability serving that scales horizontally — workers scale out under load rather than being sized in advance. Warehouse-agnostic on open-source Malloy, and the model is portable because the language is
dbt: Metric queries served through the hosted dbt Cloud Semantic Layer

What you pay for

Credible: Usage, not seats — unlimited users on every plan, so adding people never changes the bill. Three meters, pooled per organization and starting free: tokens, compute time and storage. Your agent’s own LLM tokens are never billed to you, and a query that runs on your own warehouse is not metered for the compute — only for the result it hands back
dbt: Per-seat, plus event meters — models built, and metrics queried through the Semantic Layer

	Credible	dbt
Category	The AI Analytics Engine — you write down what data means, and it generates the stack under it	Transformation framework with an add-on metrics layer
Who it is built for	The whole organization, not one central team — ops, finance and product; spreadsheet users through ML engineers	Analytics engineers who own the transformation layer
Primary interface	The agent you already use, over MCP — plus Credible Workspaces, dashboards, notebooks and reports for the people who want a UI. Every one of them reads the same model, and the model lives in your repository	Code — models authored in SQL, Jinja and YAML through an IDE or CLI, with BI tools on top
Modeling language	Malloy — a modern programming language for data: imports, inheritance, public and private members, and queries that compose into new sources	SQL with Jinja templating plus YAML, and MetricFlow for metrics
What is open	Open core. Malloy is open source, and we build and maintain Malloy Publisher — the open-source server for Malloy models, and where our agent skills and MCP retrieval tools are open source too, runnable in any agent with no Credible account. Your model is code in your repository, and Credible is in the Apache Ossie ecosystem for semantic interchange	Open core — dbt Core and MetricFlow are Apache 2.0, the platform is commercial, and MetricFlow is a reference implementation for Apache Ossie
How the model gets built	Model first: you write down what data means, and the engine derives the pipelines, storage and serving from it. Open-source MCP tools and agent skills help you capture that meaning from where it already lives, running in your coding agent	Hand-authored models, with agent-assisted authoring
Governance	One annotation on the source, in the same file as the logic it governs — no grant objects to declare, no rule to re-attach on every surface that exposes the data, no attribute table kept in step by hand. A row filter or an authorize gate over attributes the server resolves from verified identity, and everything reading that source inherits it: dashboards, notebooks and agents alike	Lineage, tests and metric governance within the project
Materialization and caching	One annotation on the source, in the same file as the logic it materializes — no derived-table block, no rollup definitions, no refresh triggers, no orchestration run to schedule. Optional per source: query your own warehouse directly, or hold a source hot in Credible’s own storage, which is how you get fast serving without buying or banging your head against a warehouse to get it	Materializations are a core concept for transformation, but the Semantic Layer itself ships no pre-aggregation cache and compiles fresh SQL per query
How an agent finds the right data	Search. Typed targets — source, dimension, measure, view, even a dimensional value — are matched against an index of the model and come back ranked, so an agent asks for what it needs instead of picking a model and touring it	List then drill — list_metrics or get_all_models, then query_metrics or get_node_details
Where the model can be used	One model served to every surface — agents over MCP, plus APIs, an SDK, embedded dashboards, notebooks and HTML data apps. Results carry an interactive UI resource (MCP Apps, the official extension), so a client that supports it renders a real chart or table instead of the model re-narrating rows	Semantic Layer APIs into BI tools, plus an MCP server
Scale and portability	Globally distributed, high-availability serving that scales horizontally — workers scale out under load rather than being sized in advance. Warehouse-agnostic on open-source Malloy, and the model is portable because the language is	Metric queries served through the hosted dbt Cloud Semantic Layer
What you pay for	Usage, not seats — unlimited users on every plan, so adding people never changes the bill. Three meters, pooled per organization and starting free: tokens, compute time and storage. Your agent’s own LLM tokens are never billed to you, and a query that runs on your own warehouse is not metered for the compute — only for the result it hands back	Per-seat, plus event meters — models built, and metrics queried through the Semantic Layer

A different premise

Where Credible takes a different approach

These are differences in what each product set out to be, not faults in dbt. Credible was built the other way around from a BI tool: collect what your data means, write it down as a data model in open-source Malloy, and serve that meaning over MCP to the agent you already use — with a UI of our own available, but never assumed.

One language, or three that work together

Templated SQL is how a generation of data teams got their work done, and SQL plus Jinja plus YAML is a stack an enormous number of people already know — that familiarity is most of why dbt won. Its composition runs through templating, which puts real weight on convention as a project grows. Malloy is one declarative language covering transformation, modeling and materialization: where dbt persists something with a model file, a config block and an orchestration run, Credible uses one annotation on the source, in the file that already defines it.

Where caching happens

As of 2026 the dbt Semantic Layer compiles fresh SQL against your warehouse on every call rather than keeping a pre-aggregation cache of its own. That is a deliberate division of labor — dbt materializations handle it upstream in the transform, where dbt is strongest — and it means performance work happens in the pipeline rather than in the serving path. Credible puts it in the serving path, as one annotation, with the result held hot in memory rather than rebuilt by a scheduled run — so making something fast does not mean adding a step to the pipeline that produced it.

Metrics, and the context around them

MetricFlow defines metrics, defines them well, and does it in the open under Apache 2.0. Credible also captures the business rules, edge cases and trusted sources behind those metrics, with access rules attached, and delivers all of it together. That is a wider scope, and MetricFlow being open is part of what makes it reachable at all.

Transformation inside the model, not underneath it

Malloy covers transformation as well as modeling, which means a pipeline is written in the same language and the same files as the model it feeds, rather than in a separate system the model then has to be kept in agreement with. That gap is where a good deal of drift comes from — a column renamed in the transform, a measure still pointing at what it used to be called. dbt has years of depth in the corners — incremental strategies, snapshots, custom materializations, a package ecosystem — and a team who lives in those will want to keep them; Credible reads the tables dbt already builds and turns them into a governed model an agent can search and be trusted with. Both are real options, and that is the point: where transformation lives is a choice, not a handoff where one picks up after the other stops.

Other comparisons

Compare Credible with the rest

Capability claims about dbt were checked against public sources in August 2026. Products in this category change quickly — confirm anything decision-critical with the vendor. Sources: Fivetran and dbt merger — Fivetran press, June 2026; MetricFlow open source and Apache Ossie — dbt Labs, Snowflake; Semantic Layer query path — dbt Labs, unwinddata.com; dbt Projects on Snowflake and the dbt package for semantic views — Snowflake documentation, checked September 2026.

Ready when you are

Bring your own data and judge for yourself

See what a governed model looks like when agents, not dashboards, are the primary consumer.