Twelve years of field evaluation experience, from pastoralist communities in South Sudan to health facilities in rural Burundi, encoded into open-source tools.
I kept running into the same problem. You find a useful tool, you build your workflow around it, and then the pricing changes. Or the platform pivots. Or the internet goes down in the middle of a field visit and you lose access to your own work.
PRAXIS runs entirely in your browser. Everything stays on your machine. No server calls, no tracking. If the wifi cuts out in Juba or Ouagadougou, the tools still work. MIT licensed, permanently free, fully yours.
"I spent years looking for evaluation tools built by someone who had actually run an evaluation in a conflict zone. I could not find them. So I built them."
Each module runs entirely in your browser. No server, no telemetry, fully MIT-licensed. Pick any module to launch.
AI-scored recommendations across 16 evaluation designs. Accounts for data, ethics, and operational context.
Eight designs covered: RCT, Cluster, Stepped-Wedge, Diff-in-Diff. Handles ICC and attrition offline.
Visual canvas with assumptions mapping. Build your programme's logic model from inputs through to impact.
Link evaluation questions, judgement criteria, indicators, and methods into a structured exportable table.
| EQ1 | Relevance | Doc review |
| EQ2 | Effectiveness | Mixed |
| EQ3 | Impact | Quasi-exp. |
| EQ4 | Sustainability | KII + FGD |
Offline CSV/XLSX engine. Auto-detect types, summarise distributions, run regressions, flag quality issues.
298 indicators across 11 sectors. Filter by PEPFAR MER codes, sector, or keyword.
Three-layer Grant Data Evaluability Diagnostic followed by a weekly Data Quality Improvement plan. 29 implementation protocols across six dependency levels, built for grant programmes in low-resource settings.
Turn a Workbench evaluation into a live-presenter deck. Inception Brief (pre-fieldwork) and Findings Brief (post-fieldwork) share one engine: keyboard nav, speaker notes on a second screen, PDF export.
Claude Project knowledge modules encoding validated evaluation methodology. Six core modules plus sector-specific extensions covering over 20 evaluation designs and 16 validated scales.
View on GitHub →
The integrated end-to-end evaluation workflow built on the .praxis file format. From Theory of Change through to reporting, with every finding traceable back to the evaluation framework.