Large CSV Viewer & SQL Filter

Inspect CSV and TSV files safely inside your browser, filter matching rows instantly, and run local SQLite WebAssembly aggregations.

Zero-Upload Guarantee:Processed 100% locally in your browser via WebAssembly & client-side JavaScript. Disconnect Wi-Fi and it still works.(Includes a 250-row synthetic API metrics CSV sample)
Drop a .csv or .tsv file here
RFC-4180 streaming parser with automatic delimiter detection and zero server upload.

Why Open Large CSV Exports in an Isolated Browser Viewer

Opening raw telemetry exports or SaaS billing CSVs in traditional desktop spreadsheets frequently freezes the app, strips leading zeros off zip codes, or reformats ISO timestamps into locale strings. Worse, uploading customer datasets to a random online CSV viewer violates compliance boundaries.

This browser viewer parses up to the first 10,000 rows using an RFC-4180 parser that auto-detects commas, semicolons (;), and tabs (\t), keeps leading-zero text intact, and exposes both instant keyword search and a lazy-loaded sql.js WebAssembly SQL editor on table data.

Workflow & Workstation CLI Alternatives

  1. Drop a local .csv or .tsv file above, or click Try sample .csv (250 rows).
  2. Use Quick filter to grep across every column instantaneously, or export the filtered subset.
  3. Open SQL Filter (Local WASM) to aggregate metrics (for example computing average latency per API endpoint).

When scanning gigabyte-scale raw CSVs outside a browser, native workstation CLIs remain fastest:

# Filter and aggregate gigabyte CSVs cleanly from disk with DuckDB CLI
duckdb -c "SELECT endpoint, COUNT(*) AS hits, ROUND(AVG(latency_ms), 1) AS avg_ms FROM read_csv_auto('metrics.csv') GROUP BY endpoint;"

Delimiter & Encoding Notes

Frequently Asked Questions

Does opening a CSV here upload customer spreadsheets to a backend server?

No. CSV and TSV parsing, quick row filtering, and local SQLite WASM queries all run strictly inside your own browser tab memory.

How does it avoid corrupting quoted commas or multi-line text fields?

Instead of calling line.split(","), this viewer uses an RFC-4180 character-by-character state machine that honors quoted delimiters, doubled quote escapes (""), CRLF/LF line endings, and auto-detected commas, semicolons, or tabs.

Why stop loading automatically at 10,000 rows in the browser?

Stopping at 10,000 parsed rows prevents multi-hundred-megabyte logs from freezing Chrome or Safari tabs while still letting you search, aggregate via local SQL, and export filtered CSV/JSON slices instantly.

Live · Offline

JSONL / NDJSON Log Viewer

Inspect newline-delimited JSON logs and LLM fine-tuning files record by record or in a flat filterable table.

Open tool
Live · Offline

Parquet Viewer & SQL Inspector

Open .parquet files locally, inspect column schema & row groups, run SQL queries, and export CSV/JSON without uploading.

Open tool
Live · Offline

Parquet to CSV Converter

Select columns from local Apache Parquet files and export clean RFC-4180 CSV directly in browser memory.

Open tool
Live · Offline

Parquet to JSON / JSONL Converter

Convert Apache Parquet datasets to standard formatted JSON arrays or newline-delimited JSONL offline.

Open tool