Soak. Get to saturation faster

Transparent, but privacy-friendly

Soak is free to use and open to community contributions. The core system is open source and available on GitHub. Users can run analyses on their own computer or in a secure, managed environment.

Works with PDF, Word, plain text, and spreadsheet files.

Thematic analysis at scale

Upload interview transcripts, survey responses, or any qualitative text. Soak automatically generates codes and themes, complete with verified quotes and source tracking.

Works with PDF, Word, plain text, and spreadsheet files.

Researcher-controlled analysis

You write the analysis prompts -- Soak handles execution. Unlike chatbots, Soak runs predefined pipelines that give you reproducible, transparent results.

Every output traces back to specific source text.

Quote verification built in

Soak verifies that quotes actually appear in your source documents using both lexical and semantic matching. Catch hallucinations before they reach your report.

Confidence scores help you identify paraphrases vs direct quotes.

Compare analyses across runs

Run the same analysis with different LLMs to check consistency. Compare datasets -- like patients vs clinicians -- to identify shared and divergent themes.

Visualise similarities with heatmaps and similarity statistics.

Structured data extraction

Extract categorical data from text with multi-model classification. Get inter-rater agreement metrics (Krippendorff's alpha) to validate your classifications.

Support for single-choice, multiple-choice, ratings, and free-text fields.

Ready to try Soak?

Soak is currently in private beta. Request access to start analysing your qualitative data.