CLI Reference

Commands

docglow generate

Generate the documentation site from dbt artifacts.

docglow generate [OPTIONS]

Option	Default	Description
--project-dir	.	Path to the dbt project root
--target-dir	target/	Path to the dbt target directory
--output-dir	target/docglow/	Where to write the generated site
--static	off	Bundle everything into a single index.html
--theme	auto	Theme: auto, light, or dark
--title	project name	Custom site title
--select	all	Only include models matching this pattern
--exclude	none	Exclude models matching this pattern
--slim	off	Omit raw/compiled SQL from output (reduces file size 40-60%)
--ai	off	Enable AI chat panel
--ai-key	env var	Anthropic API key (or set ANTHROPIC_API_KEY)
--skip-column-lineage	off	Skip column-level lineage analysis
--column-lineage-select	all	Only analyze column lineage for this model
--column-lineage-depth	unlimited	Max hops from the selected model
--include-packages	off	Include dbt package models in lineage
--profile	off	Enable column profiling
--profile-adapter	none	Database adapter (duckdb, postgres, snowflake)
--profile-connection	none	Connection string or DB path
--profile-sample-size	10000	Max rows to sample per model
--profile-no-cache	off	Skip profile caching
--fail-under	none	Exit code 1 if health score below threshold (0-100)
--verbose	off	Enable debug logging

docglow serve

Serve the generated site locally.

docglow serve [OPTIONS]

Option	Default	Description
--dir	target/docglow/	Directory to serve
--port	8081	Port number
--host	127.0.0.1	Host address
--open/--no-open	open	Auto-open browser
--watch	off	Watch for artifact changes and auto-rebuild
--project-dir	.	dbt project root (for --watch mode)

docglow health

Show project health score and coverage metrics.

docglow health [OPTIONS]

Option	Default	Description
--project-dir	.	Path to the dbt project root
--target-dir	target/	Path to the dbt target directory
--format	table	Output format: table, json, or markdown
--select	all	Only include matching models
--fail-under	none	Exit code 1 if score below threshold

docglow mcp-server

Start an MCP server for AI editor integration.

docglow mcp-server [OPTIONS]

Option	Default	Description
--project-dir	.	Path to the dbt project root
--target-dir	target/	Path to the dbt target directory

docglow init

Generate a starter docglow.yml configuration file.

docglow init [OPTIONS]

Option	Default	Description
--project-dir	.	Where to create docglow.yml
--force	off	Overwrite existing docglow.yml

docglow profile

Run column-level profiling only (without generating a site).

docglow profile [OPTIONS]

Option	Default	Description
--project-dir	.	Path to the dbt project root
--target-dir	target/	Path to the dbt target directory
--adapter	required	Database adapter (duckdb, postgres, snowflake)
--connection	required	Connection string or path
--sample-size	10000	Max rows to sample per model
--no-cache	off	Skip profile caching
--output	target/docglow/	Output directory for profiles.json
--verbose	off	Enable debug logging

docglow publish

Publish documentation to Docglow Cloud.

docglow publish [OPTIONS]

Option	Default	Description
--token	DOCGLOW_TOKEN env var	API token
--project-dir	.	Path to the dbt project root
--target-dir	target/	Path to the dbt target directory
--no-wait	off	Don't wait for processing to complete
--verbose	off	Enable debug logging

docglow login

Authenticate with Docglow Cloud.

docglow login [--token YOUR_TOKEN]

docglow logout

Remove stored Docglow Cloud credentials.

docglow logout