dbt

Bring your dbt models. Ask better questions.

Analyze models built in dbt Cloud or your existing environment. Connect GitHub when you need isolated compilation and reviewed schema changes with Kole.

dbt · Workspace previewSample data

acme / analytics mainInteractive preview · GitHub example

Your GitHub project

Source stays with you

Isolated compilation

SQL, ready to inspect

Review in Syne

Your team decides

customer_revenue Compiled

target / compiled / customer_revenue.sqlPostgreSQL

select
  customer_id,
  sum(amount) as lifetime_revenue
from "analytics"."stg_orders"
group by customer_id

Source revision retained with this compilation

GitHub source. Model context. Human review.See the workflow

Core capabilities

The right tools for
the next question.

Explore the workspace

Existing models, governed analysis

Connect dbt Cloud metadata or import a manifest, then query built models in your PostgreSQL warehouse with Kole. No GitHub connection or runner required.

Explore the workflow

dbt

Illustrative view

acme / analytics main

dbt

Illustrative view

acme / analytics main

Built into your workflow

Analysis with a traceable source.

Connect your way

dbt Cloud, manifest, GitHub

Compilation

On demand, isolated

Changes

Human-reviewed tickets

The workflow in practice

Put dbt to work.

Explore the capabilities, working examples, and decisions that shape the way your team uses dbt.

dbt projects, including dbt Cloud analysis, manifest imports, GitHub compilation, and team-managed runners, require Syne Pro or an eligible trial.

Your dbt project already describes the data your team depends on. Bring its model definitions, declared columns, and dependencies into the same place you ask questions about the data.

Analyze the models you already build

Connect a dbt Cloud account and job with a read token, or import a manifest from your existing environment. Select the PostgreSQL warehouse where the models live. Start querying built tables and views with their definitions beside you, without connecting GitHub or setting up a runner.

Syne reads metadata from the latest successful configured dbt Cloud job when you refresh. Model rows come from your warehouse through the governed database bridge. You keep your current builds and schedules; a metadata refresh does not start a dbt job.

Your models. Your warehouse.

Latest successful run

Read model metadata on refresh. Keep builds in dbt Cloud.

Model context + Kole

Understand dependencies. Write an aggregate. Inspect the SQL.

Your PostgreSQL data

Query built tables and views through your governed connection.

Read-only queries · masking · auditGitHub and runners optional for analysis

Use Analyze data for SQL joins and aggregations, or ask Kole to help. The editor shows up to 200 rows, with masking and read-only execution. Refresh or re-import metadata when definitions change. If a refresh fails, your previous successful snapshot remains available. Model metadata is context, not a data-freshness certificate.

Read the connection guide or follow the illustrated preview walkthrough.

Keep GitHub as the source of truth

When you want compilation and schema-change review, connect a repository, branch, and project folder to an existing Syne PostgreSQL connection. Repository files and pull requests stay in GitHub. Syne reads the project and records which commit each compilation came from.

Reuse the GitHub App your team has already connected to Syne. Search its repository list, sync access after repository grants change, and select a project without entering another token. SQL pull-request reviews remain optional. Removing App access cancels affected compilation jobs; source stays in GitHub.

Compile when you need to inspect a change

Request compilation in a short-lived, isolated environment with no database credentials or network access. Inspect generated SQL and model dependencies without deploying the project. Compilation history retains the source revision and highlights resources added, removed, or changed since the previous successful compilation.

Choose a Syne-hosted runner for small projects, connect a team-managed machine for larger projects, or import a manifest from your existing dbt platform, CI, or local environment. The initial hosted runner provides 1 CPU, 512 MB memory, and up to 3 minutes 40 seconds of compilation time. Fetching the GitHub project adds separate time. If compilation reaches the limit, you can retry or import a manifest; no models are deployed.

Projects that depend on warehouse introspection or external packages can use a manifest compiled in their own dbt environment. Imported artifacts are identified as user-supplied so reviewers can distinguish them from Syne compilation results.

From source to SQL

A small space to compile. A clear view of the result.

acme / analytics

dbt_project.yml

models / stg_orders.sql

models / customer_revenue.sql

Pinned to commit 8c2e6a1

Isolated dbt compiler

Created on request · discarded after use

Compiled artifact

select customer_id,
  sum(amount) as lifetime_revenue
from analytics.stg_orders
group by customer_id

SQL + model metadata + source revision

No network accessNo database credentialsNo model deployment

Bring the compute your project needs

Connect a dedicated machine your team owns, select it in project settings, and keep compilation history and reviews in Syne. The runner connects outward over HTTPS. Its GitHub read credential stays on your machine, and each compilation runs in an isolated, offline container.

Start with 2 CPU, 4 GB memory, and a 15-minute compilation limit. Teams can configure up to 16 CPU, 16 GB, and 30 minutes per compilation. Each machine handles one job at a time. See when it is online, pause new work, cancel a job, or revoke access from Syne.

Follow the runner installation guide, including the Compose download and image-access requirements.

Your infrastructure, connected

More room for your models.

Illustrative team runner

Syne workspace

Git revision8c2e6a1

CompilationQueued → running

Status, artifacts, and review history

Agent connects outward
Jobs in · artifacts out

Your team's machine

CPU

4 GB

memory

15 min

default limit

Isolated compiler · local GitHub credential

Configure up to 16 CPU, 16 GB memory, and 30 minutes per compilation.

Team-managed execution supports larger projects while keeping the same PostgreSQL compilation scope. Builds, warehouse introspection, and production scheduling remain in your existing dbt environment. Artifacts identify the team runner that supplied them.

Query models alongside their definitions

Browse declared columns and descriptions and follow upstream dependencies. Preview an existing table or view through your Syne connection. GitHub projects also support inspecting and previewing a compiled SELECT. Single-model previews use the database bridge's read-only execution path and show up to 100 rows.

Compilation and materialization are separate. A new model needs to be built in your dbt environment before its relation can be queried.

Give Kole the context to help

Ask Kole to explain a model, inspect its dependencies, or query built models for an analysis. It uses the connected project's metadata and artifact provenance to ground its answer. For GitHub projects, it can also help draft documentation, tests, and migration SQL. Compilation requests and change-ticket creation pause for human approval.

Review schema changes with their source attached

For a GitHub project, open a change ticket from a model with explicit migration SQL, a reason, and optional rollback SQL. The ticket carries its compilation and source commit into Syne's existing review and approval workflow. A compiled SELECT is never treated as an implicit database migration. Analysis-only connections keep their builds and change process in the existing platform.

Kole + your team

From “what changed?”
to a considered decision.

Explore the model with Kole. Keep the source, proposed SQL, and reasoning together for the person reviewing the change.

ExplainDraftReview

Change ticket Illustrative example

Add context to customer revenue

customer_revenue 8c2e6a1

Proposed migration SQL

COMMENT ON COLUMN
analytics.customer_revenue.lifetime_revenue
IS 'Total order revenue per customer';

Awaiting human review

A proposal stays a proposal until your team approves it.

Production dbt builds and scheduling remain in your existing dbt environment. Syne brings the models, analysis, and review context together.

Syne-managed compilation can use a shared runner pool across teams. Available workers claim queued jobs and create a fresh isolated compiler container for each request. Teams can continue to bring their own runners for dedicated capacity.

Good to know

Frequently asked questions.

Got another question?

What can my team do with dbt?

Analyze models built in dbt Cloud or your existing environment. Connect GitHub when you need isolated compilation and reviewed schema changes with Kole.

Are these previews connected to a live database?

The examples on this page use sample data. Open SyneHQ and select your team’s configured connection to work with your own data. Controls and availability depend on your permissions and deployment.

What should I know before using dbt?

PostgreSQL preview. Analyze existing models through dbt Cloud metadata or a manifest; GitHub and runners are optional. Queries run against your connected warehouse. Offline compilation does not build models, perform warehouse introspection, download packages, or certify tests.