Skip to main content

Integrating Data

ToolJet AI can read and understand your existing database schema to build applications with queries and data bindings already configured. Internal tools are often built to solve the problem of scattered data, information spread across multiple databases, APIs, and services. By connecting your real data sources, you can build on top of your existing data and create production-ready applications with AI rather than just prototypes with mock data.

How It Works​

When you provide a prompt, ToolJet AI asks whether you want to use your existing data or proceed with sample data. If you choose existing data, it will ask you to select your data source and then AI will be able to read through your database tables or API endpoints to identify the entities relevant to your application. It then presents an entity mapping for your approval before generating the final application.

A typical flow looks like this:

Prompt → Spec Doc → Select Data Source → Entity Mapping → App Generation

The exact sequence can vary based on the kind of prompt you provide or how you progress through each step.

Data Privacy
  1. ToolJet AI does not access your data, it only uses the table schema for generating applications.
  2. ToolJet AI cannot delete tables or data, even if those tables were originally generated by the AI itself.

Building With Existing Data​

Starting With Existing Data​

  1. Enter a prompt: Describe the application you want to build in the prompt input on the dashboard.
  2. Choose your data: The AI will ask if you have an existing data source or want to proceed with sample data. Select existing data.
  3. Review the spec doc: The AI generates a specification document outlining the features, navigation, and requirements. Review and approve it.
  4. Select data source: Choose data source needed for your application. You can select only one data sources.
    Choose Your DS
  5. Review entity mapping: The AI reads your database tables and API endpoints, then presents a mapping of entities to the relevant tables. Review and approve the mapping.
    Entity Mapping
  6. App generation: The AI generates the final application with all queries and data bindings configured against your real data.

Starting With Sample Data and Connecting Later​

If you start with sample data, you can connect your existing data sources at any time:

  1. Generate with sample data: Build your application using sample data first.
  2. Prompt to connect data: Enter a prompt in the AI chat asking to connect your application to real data.
  3. Select data source: Choose the data source you want to connect.
  4. Review entity mapping: The AI maps your application's entities to the relevant tables. If there's missing information, the AI will ask whether it should update the schema or create new tables as needed.
  5. Approve and regenerate: Once you approve the mapping, the AI regenerates the application with real data bindings.

Entity Mapping​

Entity mapping is the step where the AI shows you which database tables or API data sources it will use for each entity in your application.

  • For databases — The mapping shows the specific table names that will be used for each entity.
  • For APIs (OpenAPI) — The mapping shows the data source name.

An entity can use more than one table. For example, an Orders entity might pull from both an orders table and an order_items table.

You can modify the entity mapping before approving it:

  • View the complete list of tables available in your database.
  • Search for a specific table by name.
  • Add or remove tables for any entity.

Limitations​

  • Database support is currently limited to PostgreSQL and MongoDB.
  • API support is limited to OpenAPI specifications.
  • You can only connect to one data source at a time.

Need Help?​