Code-native data transformation — SQL and Python models, no Jinja required.
Run models, preview data, view lineage, and plan changes without leaving VS Code.
Features
- Run / Test / Build / Preview / Compile / Docs / Lineage — CodeLens buttons above each model file
- Model tree view in the sidebar — models grouped by materialization type (View / Table / Incremental / Snapshot / Python)
- Interactive DAG — visual dependency graph with clickable nodes, color-coded by materialization type
- Rich lineage view — toggle between card list and interactive SVG canvas showing upstream/downstream dependencies
- Model documentation — browse column names, types, descriptions, and tags from schema YAML
- Model source — view compiled SQL in a side tab
- 37 SQL snippets — surrogate_key, safe_divide, datediff, haversine, and more
- PII scan — detect sensitive columns in any model
- Open Studio — launch the browser dashboard (kelpmesh studio) from the command palette
Requirements
Install the KelpMesh CLI:
pip install KelpMesh
Quick Start
- Open a folder containing a kelpmesh.yml project (or run kelpmesh init)
- Open any .sql file in models/ — you'll see Run / Test / Preview buttons above it
- Use the sidebar tree to browse all models
- Run KelpMesh: Plan (dry run) to see what would change before running
Configuration
| Setting | Default | Description |
|---|---|---|
| kelpmesh.pythonPath | "" | Python executable with KelpMesh installed. Blank = workspace interpreter. |
| kelpmesh.projectDir | "" | Path to project root (folder with kelpmesh.yml). Blank = workspace root. |
| kelpmesh.autoRunOnSave | alse | Auto-run model on save. |
| kelpmesh.showCodeLens | rue | Show Run / Test / Preview buttons above model files. |
License
Apache 2.0