# Shaper for AI Agents

<Aside title="Work in Progress">
This functionality is still experimental. We are currently figuring out what the best way is to support different workflows and integrate agents.
</Aside>

_Shaper is a great way for agents to query and visualize data.
Since everything is represented as SQL, it's easy to verify the correctness of the queries that power visualizations._

Use Shaper's API to create temporary dashboards, query data and generate images or PDFs directly from SQL. Learn how to use the API by visiting the API docs:
https://taleshape.com/shaper/docs/api-endpoints.md

To understand how to build dashboards using SQL, see the SQL reference:
https://taleshape.com/shaper/docs/dashboard-sql-reference.md

To understand the available data start by running this query to list all tables:
```sql
SELECT
  concat(database_name, '.', schema_name, '.', table_name) AS name,
  estimated_size,
  column_count,
  comment
FROM duckdb_tables()
WHERE NOT internal AND NOT temporary;
```

Then you can run follow-up queries to get more details about the columns in relevant tables:

```sql
SELECT
  column_name,
  data_type,
  is_nullable,
  column_default,
  comment
FROM duckdb_columns()
WHERE NOT internal
  AND concat(database_name, '.', schema_name, '.', table_name) = 'your_table_name';
```