How to query

1. 1Open one or more JSON files in the editor — each becomes a SQL table with a sanitized name.
2. 2Navigate to the SQL Workspace. Your tables appear in the sidebar with column names and row counts.
3. 3Write a SQL query in the editor (e.g., SELECT * FROM users WHERE age > 25).
4. 4Press Run or Ctrl+Enter. Results render as a table below the query editor.
5. 5Export results as JSON, CSV, or XLSX, or open them as a new editor tab.

Under the hood

• `DuckDB` WASM engine — The full DuckDB analytical database compiled to WebAssembly. Supports window functions, CTEs, CASE expressions, UNION, subqueries, and most of the SQL:2016 standard.
• Automatic table registration — Each open editor tab with valid JSON is registered as a DuckDB table using registerTable(). Table names are sanitized to valid SQL identifiers with collision avoidance.
• `BigInt`-safe serialization — DuckDB can return BigInt values for large integers. The export pipeline maps BigInts to Numbers to ensure JSON.stringify compatibility.
• Multi-format export — Results can be downloaded as JSON, CSV (with proper RFC 4180 escaping), or XLSX with spreadsheet-formula escaping.
• Add external tables — Upload additional JSON files directly into the SQL workspace without leaving the page. They're registered as new tables and also persisted as editor tabs.

Real-world scenarios

• Ad-hoc analytics on API data — Pull a JSON response from an API, load it as a table, and run GROUP BY / COUNT(*) queries to summarize without spinning up a database.
• Cross-file JOINs — Open users.json and orders.json in separate tabs, then JOIN them on user_id to correlate data across files.
• Data filtering and transformation — Use WHERE clauses, CASE expressions, and string functions to extract exactly the rows and columns you need, then export as CSV.
• Quick aggregation — SUM, AVG, MIN, MAX, COUNT — run aggregate functions on JSON arrays without writing JavaScript.

Sample query

-- Given a "users" table from your JSON tab:
SELECT
  city,
  COUNT(*) AS user_count,
  ROUND(AVG(age), 1) AS avg_age
FROM users
WHERE active = true
GROUP BY city
ORDER BY user_count DESC;

-- Result:
-- +----------+------------+---------+
-- | city     | user_count | avg_age |
-- +----------+------------+---------+
-- | New York |          5 |    31.2 |
-- | London   |          3 |    28.7 |
-- | Tokyo    |          2 |    34.0 |
-- +----------+------------+---------+

Questions

Related tools