Changelog

2026-05-11 · 20 new tools — competitive parity push

Big batch. The site now has 35+ tools. Highlights below.

• New conversions. JSONL ↔ YAML (multi-doc separated by ---), JSONL ↔ TSV (tab-separated, for BigQuery / Snowflake / awk), JSONL ↔ XML (configurable root / item tag), and JSONL ↔ Excel (XLSX) — a real .xlsx workbook built in-browser with a tiny ZIP writer, no SheetJS dependency, no upload.
• Transform / query batch. jq-style query playground with path access, pipes, select, map, sort_by, group_by, object construction, and the rest of the day-to-day jq toolkit, running entirely in the browser. Plus no-code filter, sorter, sampler (random reservoir, head/tail, every-Nth, stratified), find & replace (string or regex, scoped or raw), flatten / unflatten.
• JSONL diff. Semantic diff by key path (the default), by line position, or by canonical-object hash, with per-field diff for changed records.
• Fine-tune validators per vendor. The all-in-one validator is gone — replaced by dedicated pages with vendor-specific samples, error messages, and FAQs: OpenAI, Anthropic, Gemini, Llama / ShareGPT, Mistral.
• Fine-tune workflow. Visual dataset editor with one-click export to any vendor shape. Train/val/test splitter with seeded reproducibility and optional stratification. System-prompt deduplicator for finding duplicate user prompts that inflate training cost.
• Nav. Megamenu expanded to four full categories — Convert, Inspect, Transform, AI fine-tuning — with every new tool linked.

2026-05-09 · Fine-tune validator: Google Gemini support

• The LLM fine-tune validator now covers four formats — pick Google Gemini (contents + parts) from the dropdown.
• Validates Vertex AI's supervised tuning shape: optional systemInstruction, a contents[] array starting with a user turn, strict user/model alternation (note: "model", not "assistant"), and at least one part with text / inlineData / fileData in every entry.

2026-05-07 · Home page redesign

• Tool first. The auto-fixer is now embedded directly on the home page so you can paste a broken JSONL line and see it repaired without an extra click.
• Full-width layout. Removed the fixed-width container so the tool panes can use the whole window.
• Sidebar moved to footer. The left tool list is now a tile grid in the page footer, grouped by Convert / Inspect / Transform / Sample data.
• Link colors. Links in body copy now read as links — bold links inherit the link color instead of falling back to ink.
• Acronym annotations. Pages now spell out JSONL (JSON Lines) on first mention so first-time visitors aren't guessing.

2026-05-04 · JSONL Auto-Fixer

• Conservative repair pipeline: trailing commas, single → double quotes, smart quotes, NaN / Infinity / undefined, BOMs, unquoted keys, JS comments, leading garbage.
• Re-parses each fixed line — if it still doesn't pass JSON.parse, the line is reported with column and a suggested fix instead of being silently emitted.

2026-04-25 · Dataset stats panel

• Drop a JSONL file and get row count, duplicate rate, parse errors, and a per-field table with fill rate, type histogram, distinct count, and top values.

2026-04-18 · LLM fine-tune validator covers three formats

• OpenAI chat (messages), Anthropic (top-level system + alternating user/assistant), and the legacy OpenAI prompt/completion shape.
• Picker dropdown so you validate against the format you're actually training on.

2026-04-10 · Token counter

• Estimates total tokens per file and projects fine-tune cost per model so you can budget the run before submitting.

2026-03-22 · JSONL ↔ Parquet

• Convert JSONL to Parquet (and back) entirely in the browser using DuckDB-WASM. Compressed columnar output ready to drop into a warehouse.

2026-03-05 · Schema inferrer

• Walks every record and emits a Draft-07 JSON Schema with required fields, union types, and enum detection.

Earlier

Initial release: validator, viewer, formatter, deduplicator, splitter / merger, and converters for CSV ↔ JSONL and JSON ↔ JSONL. All 100% client-side.