SAQL Editor — a browser playground for SQL on your files

Drop a JSON or CSV file into your browser and query it with SQL. SAQL Editor is a small web app for the SAQL engine: upload a file, write a SELECT in a real code editor (with autocomplete and live error‑checking), and get a results table back — all running locally on your machine.

It's the friendly, point‑and‑click way to use SAQL. No command line required.

▶ Try it live — no install

arjun1194.github.io/saql-editor

The whole SAQL engine is compiled to WebAssembly and runs inside your browser tab. There's no backend — your file is read locally and queried on your own machine, so your data never leaves the page. Just open the link, drop in a CSV/JSON, and write SQL.

What you get

• 📤 Upload a JSON or CSV file — it becomes a table you can query.
• ⌨️ A real code editor (Monaco — the same one that powers VS Code) with SQL highlighting.
• 💡 Smart autocomplete — it reads your file's columns and suggests table and column names (correctly typed) as you go.
• 🔴 Live error‑checking — invalid SQL is underlined in red while you type.
• 📊 Results grid — run the query and see the rows immediately.

Everything runs in‑process and on your machine — your data is never uploaded anywhere.

Run it

Easiest — the standalone binary. Download saql-web from the Releases page (macOS Apple Silicon), then:

saql-web                 # open http://127.0.0.1:7878
PORT=9000 saql-web       # …or pick a different port

The binary is fully self‑contained — the web UI is baked in, so it runs from anywhere.

From source (any platform — needs Rust and a checkout of the engine repo beside this one):

git clone https://github.com/arjun1194/saql            # the engine (sibling dir)
git clone https://github.com/arjun1194/saql-editor
cd saql-editor
cargo run                # open http://127.0.0.1:7878

How to use it

1. Open http://127.0.0.1:7878.
2. Upload a .json or .csv file — it appears in the table list with its columns.
3. Type a query, e.g. SELECT * FROM mytable LIMIT 20. Use the autocomplete as you type.
4. Hit Run to see the results.

The engine behind it speaks real SQL — JOIN, GROUP BY, aggregates (COUNT/SUM/AVG/MIN/MAX), HAVING, ORDER BY, and more. (Tip: text values use single quotes and column names with spaces use double quotes — WHERE "First Name" = 'Heather'.)

How it's built

A tiny axum web server that links the SAQL engine directly in‑process (no separate database). Uploaded files are decoded into in‑memory Arrow tables held in a shared session.

Endpoint	Purpose
POST /api/upload	Decode an uploaded JSON/CSV file into an in‑memory table.
POST /api/query	Run a SQL query, return { columns, rows } (capped at 5000 rows for the browser).
POST /api/validate	Parse‑only check that powers the live red‑underline error‑checking.
GET /api/tables	List loaded tables + their columns (feeds autocomplete).

The front‑end is plain HTML/CSS/JS with Monaco from a CDN — no build step. All three UI files are embedded into the binary, so saql-web ships as a single executable.

Two ways it runs:

1. Server mode — the saql-web binary above (Rust executes your SQL in‑process).
2. Serverless / browser mode — the engine is compiled to WebAssembly (wasm/ crate) and the page (docs/) calls it directly, with no backend at all. This is what's hosted on GitHub Pages. Rebuild that bundle with ./build-pages.sh (needs rustup target add wasm32-unknown-unknown and wasm-bindgen-cli).

Future scope

The editor grows alongside the engine's roadmap. On the editor side specifically:

• 🗄️ Connect to databases — a "Connect" dialog for PostgreSQL, MySQL, SQLite, DuckDB, so you can browse and query live tables next to your uploaded files.
• 📁 File‑system browser — open a folder of CSV/JSON files and query across them; drag‑and‑drop a whole directory.
• 🔗 GraphQL — run your saved queries through a GraphQL endpoint, and add GraphQL APIs as a data source you can SELECT from.
• 🌐 REST/HTTP sources — paste an API URL and query its JSON like a table.
• 📦 More formats — Parquet, TSV, NDJSON, Excel uploads.
• 💾 Save & share — named saved queries, shareable links, and export results to CSV/JSON/Parquet.
• 🧠 Smarter editing — semantic validation (unknown‑column errors, not just syntax), inline result previews, and query history.
• 📴 Offline / installable — cache the engine as a PWA so the playground keeps working with no network and installs like a desktop app.

✅ Already shipped: run fully in the browser — the engine is compiled to WebAssembly and the editor is hosted on GitHub Pages at arjun1194.github.io/saql-editor, zero server.

---

Built on the SAQL engine: https://github.com/arjun1194/saql