TrialCurveA WebPlotDigitizer alternative built for survival curves
Use WebPlotDigitizer for general charts. It is free, flexible and the de-facto standard for manual digitisation. Use TrialCurve when the figure is a Kaplan–Meier curve and you need the whole survival workflow: automatic tracing, the numbers-at-risk table, reconstructed IPD, hazard ratios and a validation record for HTA.
Feature comparison
| WebPlotDigitizer | TrialCurve | |
|---|---|---|
| Chart types | Any 2D chart, maps, polar, ternary | Kaplan–Meier / survival step curves |
| Multi-panel figures | One panel at a time | The whole figure: every survival panel is found and read |
| Axis calibration | Manual (click ticks) | Automatic from tick labels |
| Curve tracing | Manual, or semi-automatic by colour | Automatic: a vision model reads each curve, zooming in where needed, confirmed by pixel measurement |
| Numbers-at-risk table | Type it yourself | Read automatically, cell by cell |
| Reconstructed IPD (Guyot) | Separate tool (e.g. IPDfromKM in R) | Built in |
| HR, medians, RMST | Separate analysis | Built in |
| Validation / audit trail | Your own notes | Checks against the printed statistics, overlay, provenance, AI-use disclosure |
| API | No public API | REST API (curl, Python, R) |
| Price | Free | 3 free figures/month, then per figure |
| Time per KM figure | ~20–60 min including table and QC | About a minute for the whole figure |
When WebPlotDigitizer is the better choice
- Bar charts, scatter plots, forest plots and other non-survival figures.
- One-off extractions where you want full manual control of every point.
- Charts with unusual axes (log scales, polar or ternary plots).
Moving an existing workflow
If you already have WebPlotDigitizer CSV exports and typed at-risk tables, you can still use them: pass the at-risk table as a hint through the API and compare the reconstruction with your manual one. Many teams run both for a while and use TrialCurve as the independent second extraction.
FAQ
Is TrialCurve free like WebPlotDigitizer?
You get three figures a month free. After that it is pay-as-you-go per figure, or a team plan with included credits.
Does TrialCurve use AI?
Yes. A vision model (Claude Opus 5.5) reads the whole figure: the panels, axes, curves, numbers-at-risk table and printed statistics. Pixel measurement then checks each curve it read, and the reconstruction is checked against the printed hazard ratios, medians and event counts. If the model is unavailable, a measurement-only engine (on-device OCR and pixel tracing) takes over.
Reconstruct IPD from your own figure. Upload a Kaplan–Meier plot and get curves, numbers at risk, pseudo-IPD, hazard ratios and a validation pack. Three figures a month are free.
Try the live demo Browse trial data API docs