Comparison
Codd AI vs dbt Semantic Layer
The dbt Semantic Layer, powered by MetricFlow, lets analytics engineers define metrics in YAML inside their dbt project and serve them to BI tools and AI through APIs and an MCP server. Codd AI is an AI-native contextual semantic layer that discovers business knowledge, generates the data model and metrics for people to certify, and delivers answers through conversational analytics and agents. Choose dbt if your analytics engineers want metrics as code next to their transformations; choose Codd AI when you need business context for AI without hand-writing every definition.
Last reviewed October 2026, based on public documentation. Spotted something out of date? Let us know.
At a glance
| Capability | Codd AI | dbt Semantic Layer |
|---|---|---|
| Core approach | AI-native contextual semantic layer: business knowledge, rules and metrics in one governed layer | Metrics layer inside dbt, powered by MetricFlow |
| How metrics are defined | AI generates the ontology, data model and metrics; people review and certify | Written in YAML (semantic models, entities, dimensions, metrics) in the dbt project, via Git |
| Who maintains it | Data teams and business owners together, with certification workflows | Analytics engineers, through pull requests |
| Business knowledge beyond the database | Built in: captures definitions and rules from documents, BI reports and experts | Metric logic only; descriptions live in YAML |
| BI and spreadsheets | Power BI, Tableau, Looker, Excel | Power BI, Tableau, Google Sheets, Excel, Hex, Mode, Sigma (preview) and more; JDBC, GraphQL and Python SDK |
| AI agents and LLMs | Conversational analytics built in, plus MCP, Slack and Teams | dbt MCP server for AI tools such as Claude and Cursor |
| Repeatable analysis and automation | Playbooks and Actions | Not included |
| Open source | Commercial platform | MetricFlow announced as open source (Apache 2.0) in 2025; hosted APIs require a paid dbt plan |
What is dbt Semantic Layer?
The dbt Semantic Layer lets teams define business metrics once, in the same project as their dbt transformations, and query them consistently from downstream tools. It is powered by MetricFlow, which dbt Labs announced as open source under the Apache 2.0 license in 2025.
Semantic models, entities, dimensions and metrics are written in YAML and reviewed like code. The hosted Semantic Layer APIs and integrations require a dbt Starter or Enterprise-tier plan; without one, MetricFlow can still be used from the command line.
Native integrations include Power BI, Tableau, Google Sheets, Excel and several modern BI tools, with JDBC, GraphQL and a Python SDK for custom access. The dbt MCP server connects AI tools to the Semantic Layer.
Key differences
Metrics as code vs AI-generated, human-certified context
dbt's approach is ideal for analytics engineering teams who want metrics versioned and reviewed alongside their models. Every new metric or change is a code change. Codd AI generates the model and metrics from what your business already knows, then routes them to people for certification, so coverage grows without every definition waiting on an engineer.
Coverage for AI questions
A metrics layer answers questions about the metrics it contains. Third-party reviewers note that questions outside the modeled metrics simply cannot be answered. Codd AI captures the wider business context, including rules, terminology and knowledge from documents, so AI can handle more of the questions people actually ask. See what is a semantic layer.
From answers to action
dbt focuses on defining and serving metrics. Codd AI adds conversational analytics, Playbooks that make AI reasoning repeatable, and Actions that turn insights into scheduled, human-approved execution.
Which should you choose?
Choose dbt Semantic Layer if
- Your team already runs dbt and wants metric definitions to live in the same Git workflow as your transformations.
- Analytics engineers have capacity to write and maintain every metric in YAML.
- You want an open-source metrics engine you can run yourself.
Choose Codd AI if
- You need business context for AI faster than engineers can hand-write YAML.
- Business users and owners need to review and certify definitions without working in code.
- You want conversational analytics, agents, Playbooks and Actions on one trusted foundation.
- Important definitions live in documents, BI reports and experts, not only in dbt models.
Can you use Codd AI with dbt?
Yes. Many teams keep dbt for transformations and use Codd AI on top of the same warehouse tables to capture business context and power AI analytics. dbt models are a useful input: their structure and descriptions help Codd AI generate the semantic layer faster.
Frequently asked questions
What is the difference between Codd AI and the dbt Semantic Layer?
The dbt Semantic Layer is a metrics layer where analytics engineers define metrics in YAML inside a dbt project. Codd AI is an AI-native contextual semantic layer that generates the data model, metrics and business context with AI and has people certify them, built for conversational analytics and AI agents.
Is the dbt Semantic Layer free?
MetricFlow, the engine behind it, was announced as open source in 2025. The hosted dbt Semantic Layer APIs and BI integrations require a dbt Starter or Enterprise-tier plan, according to dbt's documentation.
Can I use Codd AI and dbt together?
Yes. dbt can keep handling transformations while Codd AI adds business context and AI analytics on top of the same warehouse data.
See Codd AI on your own data
Learn how Codd AI builds a certified, contextual semantic layer in weeks, and how it fits alongside the tools you already use.