Code your interview transcripts in an afternoon, not a month

Upload your transcripts in Interviews & primary research mode, pick a framework and run the analysis. The tool codes each participant separately, attaches a quote from that participant to every code, and groups the codes into themes across your dataset. You then review every code against its quote. Your first 3 participants are free.

thematicanalysis.ai turns interview transcripts into themes in under 5 minutes. It codes each transcript separately, ties every code to a checked quote, and groups the codes into themes using Braun and Clarke's reflexive thematic analysis or another framework you choose. It is built for research analysis, so it gives you themes and evidence rather than product ideas or recommendations.

This one does. Every code carries a quote checked word for word against the transcript, and every theme lists its codes. Click a code and the transcript opens with the quote highlighted, so you can trace any theme back to who said it and in what context.

Yes. Upload interview transcripts as Word, PDF, TXT or VTT files, focus group transcripts split by speaker, or a CSV or Excel file of open-ended survey answers with one row per respondent. All three are coded the same way, so you can analyse a mixed dataset in one project.

Every quote is checked word for word against the transcript you uploaded. A code whose quote cannot be found is removed and listed in the audit trail at the end of your report, so nothing reaches your findings chapter without evidence you can open and read.

3 participants free, with full-length transcripts on every tier. The Thesis Sprint Pass ($9) covers up to 20 participants and the Advanced Project Pass ($15) up to 50. Both are one-time payments for 30 days.

Yes. When a file contains several interviews, the tool splits them into one participant each using headings or speaker labels. A single interview with an interviewer stays as one participant, and the interviewer's questions are not coded as findings.

Yes, on the Advanced Project Pass. After the tool extracts your participant characteristics table (Table 1), choose a characteristic such as role, gender, work arrangement or age. You see how many participants in each group raised each theme (for example 7 of 9 nurses vs 2 of 8 doctors), a supporting quote from each group, and the largest differences written out as sentences. Ages and other numbers are split at the median. Counts are descriptive, not significance tests.

It shows code saturation. A chart plots how many new codes each participant added, and the tool applies the base size, run length and new-information threshold method of Guest, Namey and Chen (2020) to say whether and when saturation was reached. It writes a methods paragraph with the reference, and for reflexive thematic analysis adds Braun and Clarke's (2021) caveat that saturation is a description, not proof that data collection was complete.

Yes, with a pass. The REFI-QDA (.qdpx) export is the exchange format NVivo, MAXQDA and ATLAS.ti import, with your transcripts, codes and coded passages. You can also export CSV codebooks. The Word report is free.

Your transcripts are not stored on our servers, and the AI provider does not use data sent through its API to train its models. Your project is saved in your own browser. Remove names and identifying details before uploading, as your ethics approval will usually require.