Deploy dbt-tools

dbt-tools reads dbt artifacts from a target root. The same target root configuration works across CLI, Web, and MCP.

Supported target roots

Type	Format	Example
Local directory	Filesystem path	./target
Amazon S3	s3://bucket/prefix	s3://my-bucket/dbt/prod/latest
Google Cloud Storage	gs://bucket/prefix	gs://my-bucket/dbt/prod/latest

How to specify the target root

Use the --dbt-target flag or the DBT_TOOLS_DBT_TARGET environment variable:

# Flag
npx @dbt-tools/cli status --dbt-target ./target --json

# Environment variable
export DBT_TOOLS_DBT_TARGET=./target
npx @dbt-tools/cli status --json

All three surfaces (CLI, Web, MCP) accept the same --dbt-target flag and environment variable.

Required artifacts

Every target root must contain:

File	Required
manifest.json	Yes
run_results.json	Yes (for run-related commands)
catalog.json	No (enables column-level metadata)
sources.json	No (enables source freshness data)

Deployment guides

• Local target directory — use ./target from a dbt project
• S3 — configure credentials and read artifacts from an S3 bucket
• GCS — configure credentials and read artifacts from a GCS bucket
• GitHub Actions — run dbt-tools in CI
• Credentials — credential precedence and least-privilege setup

Related

• Configuration reference — full list of environment variables
• Local and remote artifacts — how remote roots work
• Trust & Safety — data boundaries and production hardening