Skip to main content

Software guide

Thematic Analysis Software: What to Use and What to Look For

Thematic analysis software helps you code your data, gather the coded extracts behind each candidate theme, and keep track of how your themes change as you analyse. You can do thematic analysis without software, and many researchers start that way. But once you have more than a handful of interviews, finding every extract for a code, merging and renaming codes, and keeping a record of how your themes developed all become slow and error-prone by hand.

This guide covers what thematic analysis software should do at each phase, why reflexive and codebook thematic analysis need different things from a tool, the software researchers actually use, and how to use AI without handing over your interpretation. For prices and a side-by-side comparison of ten tools, see our guide to qualitative data analysis software.

What the software should do at each phase

Braun and Clarke's six phases are a useful way to judge any tool, because each phase asks something different of it. The phases are recursive, so the software needs to let you move back and forth easily.

  1. 1Familiarisation — read transcripts comfortably in one place and write notes or memos as you go, linked to the passages that prompted them.
  2. 2Coding — label passages quickly, apply more than one code to the same passage, and add new codes freely as ideas emerge.
  3. 3Generating initial themes — group codes into candidate themes and see every coded extract that sits behind each one.
  4. 4Developing and reviewing themes — merge, split, rename and regroup codes and themes without losing the extracts, and check each theme against the full dataset.
  5. 5Refining, defining and naming themes — write a definition for each theme alongside its extracts, and keep a record of how and why names and boundaries changed.
  6. 6Writing up — export themes, definitions and well-chosen extracts in formats you can work with, plus the record of your analytic decisions for your methods chapter or appendix.

Reflexive vs codebook thematic analysis: different software needs

Thematic analysis is a family of methods, and the version you use changes what you need from a tool. Name your version in your methods chapter, and choose software that does not push you into another one.

  • Reflexive thematic analysis: coding is open and evolves as you analyse, there is no fixed codebook, and agreement between coders is not the aim. Look for flexible coding, easy renaming and merging, memos for reflexive notes, and a history of how your codes and themes changed. You do not need inter-coder agreement tools.
  • Codebook thematic analysis (including template and framework analysis): a structured codebook, often partly developed in advance, organises the analysis. Look for code definitions with inclusion and exclusion rules, a hierarchy you can export, and, for team projects, shared coding and ways to compare coders.
  • Coding-reliability thematic analysis: several coders apply a fixed codebook and agreement between them is measured. Look for inter-coder agreement statistics such as Cohen's kappa; NVivo, ATLAS.ti, MAXQDA and Dedoose all offer them.

The software researchers use for thematic analysis

Most qualitative data analysis software can support thematic analysis. The differences are in cost, learning curve, collaboration and how each tool handles AI.

  • NVivo, ATLAS.ti and MAXQDA: full-featured desktop packages, often free through universities. Powerful for large datasets and team projects, but with a learning curve.
  • Delve: a browser-based tool designed to be easy to learn, popular with researchers doing thematic analysis for the first time, with AI-assisted coding.
  • Quirkos: simple and visual, one of the cheapest paid options, and deliberately free of AI analysis.
  • Taguette and QualCoder: free and open-source. Taguette is the simplest; QualCoder does more, including memos and reports.
  • QualIntel OS: AI-assisted software built for academic research, with separate modes for reflexive and codebook thematic analysis. The AI surfaces evidence and you make every decision, all recorded in an audit trail. (We make it.)

Can AI do thematic analysis?

AI can help with parts of thematic analysis: surfacing passages that may be relevant to a code, suggesting codes for you to consider, and pulling together the extracts behind a candidate theme. What it cannot do is the interpretive work that reflexive thematic analysis is built on. A theme is a pattern of shared meaning around a central idea that you construct; a summary of what participants talked about is a topic, not a theme. AI tools are very good at producing topic summaries, which makes them easy to mistake for themes.

If you use AI, keep it in a supporting role. Make every coding and theme decision yourself, keep a record of what the AI suggested and what you did with it, and disclose its use in your methods section. Check your institution's AI policy and your ethics approval first, including whether participant data may be sent to an AI service.

A checklist before you choose

  • Does your university already license a tool? Start there.
  • Does the tool fit your version of thematic analysis, or push you towards a fixed codebook you do not want?
  • Can you merge, split and rename codes and themes without losing extracts, and is there a record of those changes?
  • Can you write memos linked to codes, extracts and themes?
  • If you plan to use AI: does the tool record every suggestion and your decision on it?
  • Where is your data stored, and is it used to train AI models? Check this against your ethics approval.
  • Can you export themes, definitions and extracts for your write-up, and the audit trail for your appendix?

How QualIntel OS supports thematic analysis

QualIntel OS has separate modes for reflexive and codebook thematic analysis, so its guidance and reporting follow the version you are using. The AI surfaces candidate evidence and suggests codes; you accept or reject each one, and nothing is coded automatically. You can rename, merge and regroup codes and themes as your analysis develops, and every change is recorded in an audit trail you can show your supervisor or examiners, along with an AI-disclosure statement for your methods chapter.

Frequently asked questions

What is the best software for thematic analysis?

It depends on your version of thematic analysis, your budget and whether your university provides a licence. NVivo, ATLAS.ti and MAXQDA are full-featured and often free through universities. Delve and Quirkos are easier to learn. Taguette and QualCoder are free. If you use AI and need to show examiners how you reached your themes, choose a tool that records every AI suggestion and your decision on it.

Do I need software for thematic analysis?

No. You can do thematic analysis by hand or in Word or a spreadsheet, and many researchers do for small projects. Software helps as your dataset grows: retrieving every extract for a code, merging and renaming codes, and keeping a record of how your themes developed all take much longer by hand.

Is there free thematic analysis software?

Yes. Taguette and QualCoder are free and open-source, and many universities give students free access to NVivo, ATLAS.ti or MAXQDA. Some AI-assisted tools, including QualIntel OS, also have free plans.

Can I do thematic analysis in NVivo?

Yes. NVivo is widely used for thematic analysis: you code your data to nodes (codes), group them into candidate themes, and use queries to review the extracts behind each theme. The same is true of ATLAS.ti and MAXQDA.

Can ChatGPT do thematic analysis?

It can summarise topics and suggest codes, but it cannot do the interpretive work that thematic analysis requires, and its summaries are easy to mistake for themes. General chatbots also keep no record of what they suggested, which makes AI use hard to report honestly. If you use AI, keep every decision yourself, record its suggestions, and check your institution's AI policy and ethics approval first.

Does thematic analysis software calculate inter-rater reliability?

Many tools do, including NVivo, ATLAS.ti, MAXQDA and Dedoose. Only coding-reliability approaches need it, though. Reflexive thematic analysis does not use inter-rater reliability, because it assumes there is no single correct coding of the data.

Related guides

Run your thematic analysis in QualIntel OS

AI surfaces candidate evidence, you confirm every decision, and the audit trail writes itself. One project free, no card.

Start for free