UtilityKit

500+ fast, free tools. Most run in your browser only; Image & PDF tools upload files to the backend when you run them.

SQL Online Runner

Query CSV and JSON files with SQL in your browser — powered by DuckDB-WASM.

About SQL Online Runner

SQL Online Runner lets you run SQL queries against CSV and JSON files directly in your browser, powered by DuckDB-WASM — a full analytical SQL engine compiled to WebAssembly. Upload one or more files and each is automatically registered as a virtual table named after the filename. Then write any SQL query — SELECT, GROUP BY, JOIN, window functions, aggregates — and see paginated results instantly. Export query output as a CSV file with one click. DuckDB supports a broad ANSI SQL superset plus analytical extensions like window functions, PIVOT, list comprehensions, and STRUCT types. The first use downloads DuckDB-WASM (~10 MB) which is cached by the browser. Your files are processed entirely in memory inside your browser tab and are never uploaded to any server.

Why use SQL Online Runner

  • Full DuckDB SQL engine — window functions, JOINs, aggregates, PIVOT all work
  • No upload — files are processed entirely in your browser tab
  • Supports both CSV (auto-detected schema) and JSON (newline-delimited or array)
  • Multi-file support — upload several files and JOIN between them
  • Paginated results with one-click CSV export
  • Full DuckDB SQL engine — window functions, JOINs, aggregates, PIVOT/UNPIVOT, CTEs all work as in a desktop database

How to use SQL Online Runner

  1. Click the file input and upload a CSV or JSON file.
  2. The file is registered as a virtual table using the filename (without extension) as the table name.
  3. Write a SQL query in the editor referencing the table name.
  4. Click Run Query — DuckDB loads on first use (~10 MB) then executes instantly.
  5. Browse paginated results and click Export CSV to download the output.
  6. Click the file input and upload one or more CSV or JSON files (drag-drop also works).
  7. Each file is registered as a virtual table using the filename (without extension) as the table name — see the table list above the SQL editor.

When to use SQL Online Runner

  • Filtering or aggregating a large CSV without opening Excel or writing a Python script
  • Exploring the shape and contents of a JSON API response saved to a file
  • Joining two CSVs on a common key without a database server
  • Checking data quality (NULLs, duplicates, outliers) with SQL GROUP BY queries
  • Teaching SQL concepts with real data files in an accessible browser environment
  • Joining two CSVs on a common key without spinning up a database server

Examples

Top 5 categories by revenue

Input: SELECT category, SUM(amount) AS revenue FROM sales GROUP BY category ORDER BY revenue DESC LIMIT 5;

Output: category | revenue ---------|-------- electronics | 41250.00 books | 18900.50 apparel | 15240.75 grocery | 9120.25 beauty | 7430.10

JOIN two CSVs on user_id

Input: SELECT u.name, COUNT(o.id) AS orders FROM users u LEFT JOIN orders o ON o.user_id = u.id GROUP BY u.name ORDER BY orders DESC;

Output: name | orders ------|------ Ada | 12 Grace | 9 Bob | 4 Linus | 0

Window function: 7-day moving average

Input: SELECT day, sales, AVG(sales) OVER ( ORDER BY day ROWS BETWEEN 6 PRECEDING AND CURRENT ROW ) AS ma7 FROM daily_sales ORDER BY day;

Output: day | sales | ma7 -----------|-------|------ 2026-01-01 | 100 | 100.0 2026-01-02 | 120 | 110.0 2026-01-03 | 90 | 103.3 ... (continues)

Tips

  • Filenames with spaces or hyphens become table names with underscores — quote them as "my_file" if needed.
  • Use 'DESCRIBE table_name' to inspect inferred column types before writing queries.
  • DuckDB's read_csv_auto handles tricky CSVs (quoted commas, mixed types) automatically — no schema needed.
  • Window functions (ROW_NUMBER, RANK, LAG/LEAD) are first-class — use them for time-series and top-N-per-group queries.
  • For JSON files, use json_extract(col, '$.path') to dig into nested structures within a query.
  • Wrap multi-statement scripts in BEGIN; ... COMMIT; if you need transactional semantics inside the browser session.
  • If memory pressure shows, drop unused tables with DROP TABLE name; the registered virtual tables stay until reload.

Frequently Asked Questions

What SQL dialect does DuckDB use?▾
DuckDB supports a broad subset of standard SQL plus analytical extensions including window functions, list comprehensions, PIVOT/UNPIVOT, and STRUCT types.
Is my data uploaded to a server?▾
No. DuckDB runs entirely in your browser as a WebAssembly module. Your files are loaded into browser memory and never transmitted anywhere.
What file formats are supported?▾
CSV and JSON. For CSV, DuckDB auto-detects delimiters and column types. For JSON, it reads newline-delimited or array-wrapped records.
Is there a file size limit?▾
Limited only by browser memory. Files up to around 100 MB generally work well; very large files may slow down or exhaust available memory.
Can I query multiple files with a JOIN?▾
Yes — upload multiple files and each is registered as its own table. Reference them by their respective table names in a JOIN query.
How do I reference the table name in a query?▾
The table name is the filename without the extension, with non-alphanumeric characters replaced by underscores. The registered table names are shown above the SQL editor after upload.

Explore the category

JSON Path Finder Query JSON values using JSONPath-style paths Graphviz DOT Renderer Render Graphviz DOT language graphs in-browser with WASM — supports dot, neato, circo, fdp, sfdp, and twopi engines. XLSX to CSV Converter Convert Excel XLSX or XLS files to CSV in your browser. Pick sheet, delimiter, preview, and download. CSV to XLSX Converter Convert CSV files or pasted text to Excel XLSX in your browser. Auto-detect delimiter, preview, set sheet name, download. Audio Compressor Shrink MP3, WAV, M4A, OGG and FLAC files by re-encoding at a lower bitrate. Choose 32–320 kbps, pick MP3, OGG or M4A output, then download. 100% browser-side via ffmpeg.wasm. GIF to Video Converter Convert animated GIFs to MP4 (H.264) or WebM (VP9) video files. Smaller, smoother and far more efficient than GIF. Browser-only via ffmpeg.wasm.
How SQL Query Planners Decide What Your Query Does You write SQL declaratively, but the database engine decides how to execute it. We open up Postgres EXPLAIN to show how query planners actua May 2026 · 9 min SQL Basics: From Zero to Writing Real Queries SQL is one of the most valuable skills a developer can have — and the fundamentals are surprisingly approachable. Start here to write your f Mar 2026 · 12 min NoSQL vs SQL: Choosing the Right Database for Your Project The choice between SQL and NoSQL is not about which is better — it is about matching the data model to your access patterns. Here is a frame Mar 2026 · 10 min

Glossary

DuckDB
An in-process analytical SQL database (similar in spirit to SQLite but optimized for OLAP workloads), written in C++ and compiled to WebAssembly for browser use.
DuckDB-WASM
The official WebAssembly build of DuckDB that runs entirely in a browser tab via Web Workers, with no native dependencies.
WASM (WebAssembly)
A portable binary instruction format that runs in browsers at near-native speed, used here to execute DuckDB's C++ code client-side.
SQLite
Another popular in-process SQL database; DuckDB targets analytical (OLAP) queries while SQLite targets transactional (OLTP) workloads.
Window function
A SQL construct (ROW_NUMBER, RANK, LAG, LEAD, SUM OVER) that computes results across rows related to the current row without collapsing them via GROUP BY.
CTE (Common Table Expression)
A WITH-clause query that defines a named temporary result set used in a subsequent SELECT, often used to organize complex queries.
PIVOT/UNPIVOT
DuckDB extensions that reshape rows into columns (PIVOT) or columns into rows (UNPIVOT), useful for cross-tab reports.
Virtual table
A table that exists only in DuckDB's session — backed by a CSV or JSON file in this tool — and disappears when the page reloads.
REPL
Read-Eval-Print Loop. This tool acts as a SQL REPL where each Run Query evaluates the current editor contents and renders results below.