Torque Log Studio

Local-only web app for exploring Torque Pro (Android) CSV logs as synced time-series charts. Built with Gea, TypeScript, Vite, and uPlot.

Run

npm install
npm run dev

Open the printed URL and import a .csv export.

Install as an app (PWA)

After npm run build, serve the dist/ folder over HTTPS (or run npm run preview locally). In Chromium-based browsers, use Install app (or the menu equivalent); Safari on iOS: Share → Add to Home Screen. Install prompts require a production build — the dev server does not register the offline service worker (vite-plugin-pwa dev mode is disabled).

Supported Torque CSV assumptions

• Delimiter: comma, semicolon, or tab — inferred from the header row.
• Header row: first non-empty line is treated as column names.
• Time column: prefers headers containing time, timestamp, gps, epoch, etc.; otherwise picks the column that best parses as time/numbers on sample rows.
• Time values: numeric (seconds, milliseconds, or Unix-like), ISO-ish date strings, or HH:MM:SS(.fraction).
• Channels: columns where ≥ ~65% of rows parse as numbers are treated as plottable signals (excluding the chosen time column).
• Encoding: UTF-8 text after export; unusual encodings may mis-parse — re-export UTF-8 from Torque if needed.

Known limitations

• Very large files are downsampled for drawing (stride) so the UI stays responsive; fine detail may be thinned.
• Canvas charts are visual-first; screen-reader-friendly tabular export is not implemented in v1.
• Multi-byte row filtering: rows without a parsable timestamp are skipped so channels stay aligned.

Privacy

No telemetry and no upload: parsing runs in your browser (including a dedicated worker thread). Data does not leave your machine unless you share the file yourself.