Prompt Studio

Ask Your
Data Anything.

Type a question in plain English. Get instant reports and charts — powered by NLP. No queries needed.

Postgres_Test_DB
Quick suggestions:
Top customersRevenue trendsMonthly analysis
CustomerRevenueOrdersRegion
Global Tech Inc$124,50042North America
Nexus Solutions$98,20031Europe
Future Systems$84,90028Asia Pacific
Apex Dynamics$76,40019North America

From Question to Insight in Seconds

Type Your Question

Write in plain English. No Code knowledge needed. e.g., 'Show me monthly revenue by region for Q4'

NLP Engine Processes It

RiverGen translates your prompt into an optimized query and runs it against your connected data source instantly

Get Your Report

Results appear as a table, chart, or raw code. Download, share, or save as a template in one click

What is Prompt Studio?

The place where you tell River Gen what you want — in your own words.

A prompt can be as simple as a question or as detailed as a step-by-step instruction. Users can try different versions, compare the results, save the best prompts as templates, and organize them by workspace, project, or use case. All of this happens in a guided interface that shows exactly what is being asked and what came back.

  • Natural language to Code, instantly
  • Supports any connected database
  • Results in table, chart, or code view

Day-to-Day Use

Standardize how your team asks recurring questions.

Instead of rewriting the same instructions over and over, teams can choose a template, adjust a few options, and run it again with consistent quality. History and analytics views help them understand which prompts perform well, how often they are used, and where improvements might be needed.

  • Save best prompts as reusable templates
  • Analytics on usage and performance
  • History view of every prompt run

Central Prompt Library

All your organization's prompts, versioned, tagged, and searchable.

Customer Support v3

Production
2h ago

Lead Scoring v7

Staging
1d ago

Email Personalizer

Draft
3d ago

Document Extractor v2

Production
5d ago

Result Views

Every prompt returns data three ways.

Table View
CustomerRevenueRegion
Global Tech$124kNA
Nexus Sol$98kEU
Future Sys$85kAPAC
Chart ViewRevenue by Region Q4
Code View
SELECT customer, SUM(revenue) AS total
FROM sales WHERE quarter = 'Q4'
GROUP BY region ORDER BY total DESC

Connected to Your Data, Instantly

Prompt Studio works directly with your connected data sources. Switch sources mid-session without losing your work.

PostgreSQL

Connected

MySQL

Available

Snowflake

Available

CSV Upload

Available

Universal Query Execution

Switch data sources without rewriting prompts — just select and re-execute to see how different data sets compare.

Centralized Credentials

All sources managed centrally in the Data Sources product — credentials stored once, reused everywhere across your workspace.

Contextual History

Prompt history is linked to its source so you always know exactly what data backed each result and when it was run.

Save, Reuse, and Track Everything

Templates

Monthly Revenue Report

Top customers by revenue, grouped by month and region

Churn Risk Analysis

Identify customers with declining order frequency over 90 days

Product Performance

Compare units sold and revenue across all product categories

History

Top 10 customers by revenue this quarter

Postgres_DB2m ago

Monthly churn rate last 6 months

Snowflake_Analytic45m ago

Revenue by region Q4 2024

Postgres_DB3h ago

New signups by channel this week

MySQL_Prod1d ago

Average order value by product

CSV_Upload2d ago

Understand How Your Team Uses Data

Track which prompts run most, who runs them, and how results change over time with built-in analytics.

Most Used Prompt

Monthly Revenue

+23%

Run 142 times

Active Users

38

+5 this wk

Across 4 workspaces

Avg Response Time

1.2s

-0.3s

Across all runs

Data Sources Active

6

Postgres, MySQL +4

Collaboration & Governance

Leave inline comments, request reviews, and enforce approval gates before any prompt goes to production.

Team Review — customer_support_v3
SK
Sarah K.10m ago

Line 2: should we add a fallback for empty `{input}`?

AM
Alex M.25m ago

LGTM — accuracy improved +2.1% from v2.

Start Asking Questions Your Data Can Answer

Connect a data source, type your first prompt, and get a report in under 10 seconds.