Text2SQL.ai
NewText2SQL.ai turns plain-English questions into SQL, explains, fixes and optimizes queries, generates spreadsheet formulas and regex, and can connect to MySQL, PostgreSQL or SQL Server to run read-only queries and charts.
Text2SQL.ai
Explore features, practical uses and pricing below.
Text2SQL.ai writes SQL from a plain-English request. You describe the result you want, it returns a query for your database, and you can keep asking follow-ups until the query is right. The same workspace explains unfamiliar SQL, fixes broken queries, suggests optimizations, and generates Excel or Google Sheets formulas and regular expressions. It is a fit for analysts, developers and operations people who know what they want from the data but not always the syntax.
Three ways to give Text2SQL.ai your schema
Generated SQL is only as good as the tool's knowledge of your tables, so setup matters. The connections guide describes three routes:
- Direct connection for MySQL, PostgreSQL and SQL Server, using host and credentials or a connection URL. The app tests the connection and extracts the schema.
- Extraction scripts for Oracle, SQLite, Snowflake, Redshift, BigQuery and MariaDB. You run the script in your own client and paste the output.
- AI schema parsing for anything else, such as ClickHouse, Hive, Spark or DB2, from pasted SQL scripts, migration files or a text description of the tables.
Insights, Safe Mode and the desktop app
With a live connection, the Insights view turns a question into a query, a results table, an automatic chart and a written explanation, and you can ask for a different chart type. Safe Mode is on by default for new connections and only lets read-only statements such as SELECT run; the docs suggest switching it off only for connections that need data-modifying work such as ETL or migrations. The cloud service can't reach localhost, and your database has to accept connections from the app's static IP. For local databases there's a desktop app, included with Pro, which the company says keeps credentials and data on your machine and sends only schema names to AI providers.
Answering a churn question without writing the joins
A product analyst needs monthly churn by plan for the last year. She connects the read-only replica of the company's PostgreSQL database, leaving Safe Mode on. She asks for customers who cancelled each month grouped by plan, and Text2SQL.ai returns a query joining subscriptions, plans and cancellation events. The first version counts paused accounts as churned, so she adds a follow-up excluding the paused status. She runs it through Insights, checks a few months against the billing dashboard, switches the chart to stacked bars, and saves the final SQL in her team's repository with a comment explaining the paused-account rule.
Where Text2SQL.ai needs checking
The company claims 95% accuracy on first generation for standard queries and says complex ones may need adjustment. Business definitions are the usual trap: the model can't know that your churn excludes trials unless you say so. Extraction-script and AI-parsed schemas don't connect live, so you run those queries yourself. The FAQ says processing uses OpenAI, and teams needing zero data retention are told to contact the company. If you want shared metric definitions and dashboards, compare BlazeSQL.
Trial, Pro and API access
According to the homepage FAQ, you can try every feature free for 7 days, after which a subscription is required. The getting started docs describe a Pro plan with unlimited messages, 12+ database types, team invitations and API access, plus an Enterprise plan with custom limits, private deployment and SSO. Pro includes 100 API requests a month, with more at $0.05 per request. The site doesn't publish Pro's subscription price outside the app, so check it at signup. Other tools in this group are on the AI SQL tools page.
Questions people ask about Text2SQL.ai
Can Text2SQL.ai change data in my database?
Only if you disable Safe Mode for that connection. With it on, the docs say only read-only queries can be executed.
Does it support Snowflake or BigQuery?
Yes, through extraction scripts that you run and paste, rather than a direct connection.
What does the API do?
The FAQ lists SQL generation, fixing and explanations through the API.