Skip to content

Quick Start

Get started with SQLatte in 5 minutes! This guide walks you through your first natural language queries.

Step 1: Verify Installation

Make sure SQLatte is running:

curl http://localhost:8000/health

If you see a healthy response, you're ready to go! If not, check the Installation Guide.

Step 2: Access the Interface

SQLatte provides multiple interfaces:

Open your browser to: 

http://localhost:8000/widget.html

This gives you the standalone widget interface for testing.

Add SQLatte to your HTML page:

<!DOCTYPE html>
<html>
<head>
    <title>My App with SQLatte</title>
</head>
<body>
    <h1>Sales Dashboard</h1>

    <div id="sqlatte-widget"></div>

    <script src="http://0.0.0.0:8000/static/js/sqlatte-widget.js"></script>
    <script>
        SQLatte.init({
            container: 'sqlatte-widget',
            apiUrl: 'http://0.0.0.0:8000'
        });
    </script>
</body>
</html>

If you have analytics enabled: 

http://localhost:8000/dashboard

Step 3: Your First Query

Let's try some natural language queries!

Basic Query

Type in the widget: 

Show me all customers

SQLatte will: 1. Convert your question to SQL 2. Execute the query 3. Display results in a table 4. Show the generated SQL

Filter Query

Try adding conditions: 

Show me customers from New York who signed up this year

Aggregate Query

Get some analytics: 

What's the total revenue by month for the last 6 months?

Join Query

Combine tables: 

Show me top 10 products by revenue with their category names

Step 4: Understanding Results

When you submit a query, SQLatte shows:

Generated SQL : The SQL query that was created from your natural language question

Results Table : Data returned from your database, with pagination if needed

Insights (optional) : AI-generated observations about the data

Actions : Save query, schedule it, or download results

Step 5: Save and Schedule

Save a Query

Click the Save button after running a query. Give it a name: 

Monthly Revenue Report

Schedule a Query (if enabled)

If scheduling is enabled:

  1. Click Schedule after running a query
  2. Set frequency: Daily, Weekly, Monthly
  3. Choose time: 09:00
  4. Add email recipients: team@company.com
  5. Click Create Schedule

You'll receive automated reports via email!

Understanding the Barista

SQLatte's "barista personality" means:

It remembers context : "Show me those customers again" works in follow-up queries

It suggests insights : "I notice revenue dropped in March - want to see why?"

It clarifies when needed : "Which table are you asking about - customers or orders?"

It stays friendly : Conversational tone instead of technical jargon

Common Query Patterns

Time-based Queries

Show me sales from last month
Show me daily active users for the past week
What's the trend of signups over the last quarter?

Aggregations

What's the average order value by customer segment?
Count how many orders each customer has made
Sum up revenue by product category

Rankings

Top 10 customers by total spend
Bottom 5 products by units sold
Most active users this week

Comparisons

Compare revenue between Q1 and Q2
Show growth rate month over month
Which region has the highest conversion rate?

Query Tips

Be Specific

Instead of: "Show me data"

Try: "Show me customers who made purchases in the last 30 days"

Use Time References

SQLatte understands: - "last month", "this year", "yesterday" - "past 7 days", "last quarter" - "between January and March"

Name Your Metrics

Instead of: "Calculate that thing"

Try: "Calculate total revenue", "Count active users"

Reference Table Names

If your database has multiple tables:

"Show me customers from the users table"

Example Conversation

Here's a realistic conversation with SQLatte:

You: Show me our top customers

SQLatte: Here are the top 10 customers by total order value...
[Results displayed]

You: What about last month specifically?

SQLatte: Here are the top customers for last month...
[Results displayed]

You: Can you show me what they purchased?

SQLatte: Here are the purchases from those customers...
[Results displayed with joins]

You: Schedule this to run monthly and email it to me

SQLatte: I've scheduled this query to run on the 1st of each month
and email results to your@email.com

Next Steps

Now that you've run your first queries:

  1. Explore Configuration - Customize SQLatte's behavior in Configuration Guide 2Set Up Scheduling - Automate reports with Scheduled Queries 3Enable Analytics - Track usage with Analytics Dashboard 4Customize AI Behavior - Use Runtime Prompts

Troubleshooting First Queries

"No tables found"

Your database connection might not have permissions to see tables.

Solution: Grant proper permissions or check your database configuration.

"I don't understand that question"

The query might be too vague or reference unknown tables.

Solution: Be more specific about table names and columns.

"Query timeout"

The generated SQL might be too complex or the database is slow.

Solution: Try narrowing your question or optimizing database indexes.

Results look wrong

The AI might have misunderstood your schema.

Solution: - Check table and column names match what you asked - Review the generated SQL - Be more explicit about joins and filters

Sample Questions by Database Type

Show me top selling products this month
What's the average cart value?
Which customers haven't ordered in 90 days?
Compare sales between categories
Show daily revenue for the last 30 days

How many active users do we have?
What's the churn rate this quarter?
Show signup trends over time
Which features are most used?
Calculate monthly recurring revenue

What's the conversion rate by campaign?
Show email open rates by segment
Which channels drive the most traffic?
Calculate cost per acquisition
Compare performance month over month

Show inventory levels by warehouse
Which orders are delayed?
Calculate on-time delivery rate
Show production output by shift
Identify bottlenecks in the process

Getting Help

If you get stuck:

  • Check the generated SQL to understand what query was created
  • Try rephrasing your question
  • Be more specific about table and column names
  • Open an issue on GitHub

Happy querying! ☕