The SQL Explorer feature in Max provides a powerful interface for querying data directly from a connection. Available as a tab within the Connections screen, SQL Explorer allows users to write and execute SQL queries against their connected databases. This feature is primarily used by Assistant Developers, and access is limited to users who have had the Connection shared with them.

In addition to writing SQL queries, users can select a Dataset and ask questions using natural language. Max translates the natural language prompt into SQL, enabling users to retrieve answers without needing to know SQL syntax. If the results are not as expected, users have several options to refine the query:

1. Manually Adjust the SQL: Users can directly edit the generated SQL query to fine-tune results or correct errors.

2. Modify the Natural Language Prompt: Users can rephrase or clarify their original natural language query to get more accurate results.

3. Edit the Prompt Sent to the Language Model: Users can optimize the prompt that Max's language model uses to interpret the question, giving them greater control over how queries are handled.

This feature combines the flexibility of SQL with the convenience of natural language processing, providing developers with versatile tools for working with their data.

Common Use Cases

1. SQL Power Users: Developers can write SQL queries to retrieve specific results from their datasets, bypassing natural language processing if they prefer direct control.
2. Natural Language Queries: Developers unfamiliar with SQL can ask questions in plain language, allowing Max to generate the SQL query automatically.
3. Refining Queries: Developers can fine-tune SQL, natural language prompts, or the language model's interpretation to ensure they get the precise data they need.

The SQL Explorer feature is typically used by Assistant developers rather than business users, and is only accessible to those with shared access to the relevant Connection.