How do I migrate internal tools off Retool?

Summary

• Databricks Apps let you build internal tools directly on your data without a separate low-code platform, eliminating licensing costs and vendor lock-in.
• Migration starts with auditing your Retool workflows: identify the data sources, trigger logic, and UI components, then map them to Databricks Apps.
• Databricks Apps connect natively to your Databricks workspace, so your data stays in your lakehouse and doesn't require ETL to a third-party platform.
• You can build on Python and SQL, or use pre-built Genie Agents for common tasks like data exploration, analytics, and Q&A without custom code.
• Gradual migration lets you run both platforms in parallel: build new apps on Databricks Apps while maintaining existing Retool workflows until teams are trained.

How do I migrate internal tools off Retool?
Migrating from Retool to Databricks Apps means rebuilding internal workflows on your lakehouse instead of maintaining a separate third-party platform. The migration path depends on your use case: simple dashboards can often be rebuilt faster on Databricks Apps, while complex multi-step workflows benefit from Databricks Apps' native Python and SQL support. Retool's strength is rapid prototyping for basic CRUD apps; Databricks Apps excels at data-intensive tools that need real-time access to your warehouse.

What is an internal tools migration?

An internal tools migration is the process of moving business-critical workflows from one platform to another. In this case, you're moving from Retool's visual builder to Databricks Apps, which run on your Databricks workspace and connect directly to your data. Databricks Apps have native access to your lakehouse, Unity Catalog, and governed data. This reduces complexity because your app queries your warehouse directly without needing to connect external APIs or copy data to another vendor.

Key capabilities to evaluate

Capability	Why it matters
Native data access	Your app queries your lakehouse directly without ETL or API layers, so data is fresher and queries are faster.
Data governance	Apps respect Unity Catalog access controls, row and column filters, and audit logs, so sensitive data isn't exposed to users who lack permissions.
Code flexibility	Python and SQL let you implement complex business logic, ML models, and real-time transformations that low-code builders can't express.
Team agility	Developers can version-control app code, run CI/CD pipelines, and collaborate using standard Git workflows.
Cost	Databricks Apps run on your existing compute; you don't pay per-seat licensing for an external platform.
User experience	Apps can embed Genie Agents for conversational data access, offer rich dashboards, and support interactive forms without backend coding.

How Databricks approaches internal tools

Databricks Apps are full-featured applications that run on your Databricks workspace. You build them using:
UI components and SQL. Use Node.js and React on the frontend to build forms, tables, and dashboards; connect them to SQL queries and Python code on the backend. Databricks handles authentication, compute provisioning, and permissions enforcement.
Python-based logic. Write custom Python functions to orchestrate multi-step workflows, call LLMs via Model Serving, or invoke Genie Agents for AI-powered insights. Your functions have direct access to your lakehouse via Databricks SDK.
Pre-built Genie Agents. For common tasks like "show me revenue by region" or "what are the top reasons for failed orders," Genie Agents let business users ask questions conversationally instead of clicking through a custom UI. Agents run queries against your unified data model.
Declarative Automation Bundles. Define app manifests in YAML, version them in Git, and deploy to staging and production using the Databricks CLI. Rollback and audit trails are built in.
Migration is often faster for analytics-heavy tools (dashboards, reporting, data exploration) and slowest for forms-based CRUD apps that require custom validation. For most teams, parallel operation (running both platforms during a transition window) is the safest approach.

Common migration scenarios

Retool dashboard to Databricks Apps dashboard. If your Retool app is mainly a read-only dashboard showing aggregated metrics, rebuilding it as a Databricks Apps SQL dashboard takes a few days. Both platforms can query the same data source; Databricks Apps will be faster because it queries your lakehouse directly.
Retool form + workflow to Databricks Apps form + Lakeflow Jobs. If your Retool app collects user input and triggers downstream actions, rebuild the form in Databricks Apps and use Lakeflow Jobs to execute the backend workflow. Lakeflow Jobs can call APIs, write to tables, or orchestrate multi-step processes.
Retool multi-tenant app to Databricks Apps with row-level filters. If your Retool app uses row-level security to show different data to different users, migrate to Databricks Apps with Unity Catalog row filters. Permissions are enforced at query time, so you don't need custom filtering logic in the app.

FAQs

Can I use my existing Retool workflows and queries?

No, you'll need to rewrite them. Retool's REST query builder and scripting language don't port directly to Databricks Apps. However, SQL and Python queries can often be reused with minimal changes. Start by extracting the SQL from your Retool query definitions and adapting them to your Databricks schema.

How much coding experience do I need?

For simple dashboards and forms, a data analyst with basic SQL and JavaScript can build Databricks Apps. For complex multi-step workflows, you'll want Python developers. Genie Agents let non-technical users ask questions without coding.

What about user experience and UI?

Databricks Apps support rich React components (tables, charts, forms, modals) and responsive design out of the box. You can match or exceed your Retool UI quality without extra work.

How do I handle authentication and permissions?

Databricks Apps authenticate via your Databricks workspace identity (single sign-on, SAML, SCIM). Permissions are enforced through Unity Catalog; users only see data and tables they're authorized to access, with no custom app-level ACLs needed.