TableForge is a private-by-default browser workspace for messy tabular data. Paste HTML, Markdown, CSV, TSV, or JSON; inspect and edit the result; then export it as CSV, TSV, JSON, Markdown, SQL, XLSX, or an image. Input stays in the browser—there is no application server or analytics layer.
Moving a table between a webpage, a spreadsheet, a database, and documentation is usually a trail of one-off scripts. TableForge puts that cleanup loop in one inspectable interface and keeps sensitive source data client-side.
- Four independently tested parsers normalize HTML, Markdown, CSV/TSV, and JSON into one typed table model.
- The HTML parser handles
rowspan,colspan, nested markup, links, images, hidden elements, and configurable cleanup. - The editable grid supports virtualized rendering, transpose, regex replacement, table joins, and inferred column types.
- Exports escape SQL strings and Markdown delimiters and generate real
.xlsxworkbooks in the browser. - Charts are derived from compatible numeric columns; source data never leaves the page.
- Visualization and XLSX code load only when requested, keeping the initial workspace bundle focused on parsing and editing.
npm ci
npm run checkcheck runs ESLint, the Vitest parser/export suite, TypeScript compilation, and the production Vite build. GitHub Actions runs the same gate on pushes and pull requests.
npm ci
npm run devThe production build uses a relative asset base so it can run on GitHub Pages:
npm run build
npm run previewsrc/parser/ format-specific normalization and tests
src/hooks/ browser session and derived selection state
src/components editable grid, merge, export, and visualization UI
src/exporter/ serialized and downloadable output formats
The application is intentionally client-only. That trades collaborative storage and server-side processing for a smaller privacy boundary and a deployable static artifact.
MIT — see LICENSE.