Move from Apache Superset

Rebuild a Superset dashboard in Querio using its datasets, metrics, charts and saved queries as source material.

Already exported? Go straight to Move exported work to Querio.

1. Prepare access

Install unshackle and run the commands below from your analytics Git repository.

• SUPERSET_URL with SUPERSET_USERNAME/SUPERSET_PASSWORD, or a bearer token in SUPERSET_API_KEY.
• Password login supports SUPERSET_AUTH_PROVIDER=db or ldap. SSO-only instances need a supported API identity instead.
• The source account sees only its permitted content. Role exports also require FAB_ADD_SECURITY_API = True; row-level rules on that account affect compiled chart SQL.

2. Export one report

unshackle check superset --prompt
unshackle inventory superset
unshackle migrate superset --only dashboard:DASHBOARD_ID
unshackle todo

Replace DASHBOARD_ID with the selected item's ID from inventory. The credential prompt saves values to a local, git-ignored .env; keep those values out of agent prompts.

Read migration/report.md for missing dependencies, extraction errors and unfinished translations before continuing.

Using exported files instead

Export the selected dashboards, charts or datasets as a ZIP from Superset, then:

unshackle init
unshackle import superset "dashboard_export.zip"
unshackle build superset
unshackle todo

The ZIP contains chart settings but not compiled chart SQL, roles or row-level security. Tiles can need SQL written from their dataset definitions, and filters may be listed without being connected to queries.

3. Review what needs adapting

• Virtual datasets need warehouse equivalents. Review Jinja expressions and complex metric SQL.
• Rebuild rolling, cumulative and time-comparison post-processing in SQL or Python.
• Test time-grain filters, integer division and pivot behavior. A chart saved without query context may need re-saving and re-exporting.

Save the source chart's results as CSV with dashboard filters applied. Record the source user's row-level rules and compare those same conditions in Querio.

4. Rebuild and check it in Querio

Open the repository in Claude Code, Codex or your preferred coding agent. Give it this brief:

Read https://docs.querio.ai/llms/guides/move-bi-workflow/already-unshackled.md.
Use this repository's Apache Superset export to rebuild REPORT_NAME in Querio.
Pay attention to Superset virtual datasets, Jinja, chart post-processing
and dashboard time filters.
Ask me for the target workspace, datasource and a trusted source result.

Replace REPORT_NAME with the report you chose. Follow Move exported work to Querio for CLI authentication, the full agent task, verification and sharing. Check the rebuilt Querio output against your source result before using it as a replacement.

For export settings and the full list of translation gaps, see unshackle's Apache Superset guide.