Qualitative analysis for interviews and open-ended responses
From interview recording to evidence-backed findings.
Themera transcribes your interviews, pseudonymises names and codes your material with AI. Every code comes with a verbatim quote, and anything that could not be coded stays visible. For theses, research and market research, in English and German.
- Stored and processed in the EU
- Not used for AI training
- DPA available
- English and German
Codebook
- Introduction and training12
- Changes in daily work18
- Attitude to the system9
- Wishes for the future7
Evidence
Since then the handover is simply clearer.
P1 · line 41
Quote found verbatim in the transcript
94% of responses coded · 6 without a code
One workflow
Everything from the recording to the report
Use the whole chain or just the step you need. Each part also works on its own.
The tools
What you can do with Themera
In the app
AI coding with evidence
Themera drafts a codebook with definitions and anchor examples and codes every response. Quotes that are not in the material word for word are rejected automatically.
- Kuckartz, Mayring, Braun & Clarke
- Thematic summary
- Edit the codebook and re-code
Interview transcription
Upload a recording and get a transcript with separated speakers and timestamps. Correct it while listening, then download it.
- Click a timestamp to hear the passage
- Find and replace across the transcript
- Word with line numbers, SRT, VTT, CSV
Pseudonymisation
Find names, places, institutions and contact details and replace them with labels like [Person 1]. You review every match; real names are not stored.
- Handles inflected German names
- Same person, same label across files
- Included on every plan, free too
Intercoder agreement
Re-code a random sample blind, without seeing the AI's codes. Themera compares both codings and computes Cohen's kappa per category.
- Percent agreement and kappa
- A paragraph for your methods section
- Built into every project
Coverage and data check
Before anything runs you see how your file was read. Afterwards the coverage indicator shows which responses got no code, so nothing disappears silently.
- .docx, .pdf, .txt, .csv, .tsv and audio
- Every response accounted for
- Uncoded responses listed
Exports for your report
Take results into your thesis, report or other software: a Word report, a CSV of all codings, and transcripts with MAXQDA-compatible timestamps.
- DOCX and CSV
- Transcripts as Word, TXT, SRT, VTT
- #hh:mm:ss-x# timestamps for MAXQDA
Free, no sign-up
A closer look
Every code has a quote you can check
Each code points to the exact passage it came from. Themera checks that every quote really appears in your material and drops any that do not, so you never cite a sentence nobody said.
- Verbatim evidence for every code
- Sentiment per coded passage
- You rename, merge and reassign codes
“took me another half hour every evening”
P5 · l. 18 · verified verbatim
“the handover is simply clearer”
P1 · l. 41 · verified verbatim
Transcripts that are ready for your appendix
Speakers are separated automatically. Rename them to I and P1, fix misheard terms in one go, and download a Word file with line numbers and timestamps, or subtitles for other tools.
- German and English recordings
- An email when the transcript is ready
- Delete the recording in one click
How did you experience the introduction?
It was hard at first. Then [Person 1] helped us a lot.
What changed in your daily work?
Pseudonymised before anyone analyses it
Rules catch email addresses, phone numbers and IBANs; a language model catches names, institutions and distinctive details. Nothing is replaced until you confirm it, and the mapping stays in your browser.
- Real names are never stored
- Download the mapping for your records
- Works for transcripts and free text
Mapping stays in your browser
Methods
Built around the methods you will cite
Content-structuring qualitative content analysis with main and sub-categories.
Summarising content analysis with categories developed from the material.
Structuring content analysis with theory-driven category definitions.
Reflexive thematic analysis for patterns of meaning across the data.
Who uses Themera
One dataset, one deadline: a one-off plan without a subscription.
Interviews and focus groups with an intercoder check built in.
Customer interviews, NPS comments and open survey answers at scale.
Data is stored in the EU and analysed and transcribed by Mistral AI in Paris.
Your material is not used to train AI models.
A DPA (Art. 28 GDPR) is available for organisations.
Delete recordings and projects at any time; real names from pseudonymisation are never stored.
Compared with established QDA software
| Themera | Established QDA packages | |
|---|---|---|
| Pricing model | Free tier, €69 once, or from €39/month | Licence per semester or year; often free via campus licence |
| AI coding | Codes the whole dataset, quote checked per code | Suggestions and summaries; usage limits vary |
| Transcription | Included in every plan | Available, often at extra cost |
| Getting started | First result on day one | Training pays off over several projects |
| EU data processing | Yes | Varies |
Pricing
Start free. Pay once for a thesis, or monthly for ongoing work.
Frequently asked questions
Start with your own material
Three analyses and 30 transcription minutes free, no card needed.