If you are searching for an ATLAS.ti alternative, you probably have a concrete reason: no licence through your institution, a deadline that leaves little time to learn a large program, or a coding workload that is mostly text and mostly yours alone. This article tries to help you decide honestly whether switching makes sense. It is written by someone who builds a competing tool, so read it with that in mind, and expect it to tell you in several cases to stay where you are.
First: do you need an alternative at all?
Many universities offer campus licences for ATLAS.ti or comparable programs. Ask your library or IT service before you look elsewhere. If you can use ATLAS.ti at no cost and your supervisor or team already knows it, that is a strong argument for staying: you get help when something goes wrong, colleagues can open your project, and nobody has to be convinced that your tool is suitable. ATLAS.ti also offers a free trial, so you can check whether it fits before you commit either way.
When ATLAS.ti is the better choice
- Mixed material. ATLAS.ti handles text, PDFs, images, audio, video, survey files and social media data in one project. If you need to code video scenes or photographs directly, a text-focused tool is the wrong choice.
- Memos and networks. Analytic memos that grow over months, and network views that link codes, quotations and memos visually, are central to grounded theory and many interpretive approaches. Themera has no memo or network function.
- Exploring relationships between codes. Boolean and proximity queries and code co-occurrence tables let you ask where codes overlap or follow each other. If that kind of exploration is part of your method, a workbench is built for it.
- Long-running team projects. ATLAS.ti supports several people working on one project, including on the web version. Themera can check agreement between two codings, but it is not a team project tool.
- A requirement from your institute or supervisor. If a program has been specified, use it.
When a lighter tool can be enough
The typical case looks like this: one project, a fixed deadline, you work alone, and the material is text, for example a few dozen interview transcripts or several hundred answers to open survey questions. Here the bottleneck is rarely a missing feature. It is the time it takes to learn the software and to do the first pass of assigning passages to codes.
Themera starts from the other end. You upload the material, receive a draft codebook and a first coding, and then revise both. Every code is backed by a verbatim quote, and quotes that are not exactly in the material are rejected automatically. A coverage indicator shows which responses received no code. You can edit the codebook and re-code the material, and export the result as DOCX or CSV. Audio interviews can be transcribed in German and English, and names can be pseudonymised before the analysis runs.
A decision test in three questions
- Is there a licence that costs you nothing, and do the people around you know the program? Then staying is usually the better choice.
- Do you need things a text tool does not have, such as video, images, memos, networks or co-occurrence queries? Then stay as well.
- Is your project mostly text that you have to code alone, under time pressure? Only then is a lighter tool worth trying, and even then on a small sample before you decide.
Typical mistake: choosing the tool before the method
No program decides for you whether you are doing a codebook-style analysis, reflexive thematic analysis or grounded theory. If you pick the tool first, the method often ends up shaped around it. Settle the approach first; the article on reflexive versus codebook thematic analysis helps if you are unsure which one you are actually doing. Then check which tool can carry that method.
Typical mistake: comparing feature lists instead of tasks
Feature lists are long; your task list is short. Write down what you will actually do in the next eight weeks: transcribe, pseudonymise, develop codes, code the material, support findings with quotes, and move tables and extracts into your write-up. Then check which of these steps each tool makes easier. Everything else is decoration, however impressive it looks on a comparison page.
A note on AI features and data
ATLAS.ti now has its own AI functions, including AI-assisted coding, summaries and a chat with your documents. According to its manual, its Intentional AI Coding feature works with OpenAI's GPT models, sends the document content to ATLAS.ti and OpenAI servers only after you give explicit consent, and the data is not used to train OpenAI's models. ATLAS.ti also offers features it develops itself without OpenAI. So the question is not whether a tool uses AI, but where your material goes when you use those functions, and whether that is covered by your consent forms and ethics approval.
For Themera, the answer is: data is stored and processed in the EU (Mistral AI, Paris), it is not used for training, and a data processing agreement is available. Whatever tool you use, check these points against your own project. Legal and data protection questions belong with your university's data protection office or your supervisor, not with a blog post.
What software can and cannot do here
Neither a workbench nor an AI-assisted pipeline does the interpretation for you. Software can organise material, make passages retrievable and propose a first coding. Whether a code holds up, whether two patterns belong together and what a finding means is still your decision. A useful check either way is an intercoder agreement test: in Themera you can re-code a sample blind and compare the result using Cohen's kappa. The guide on coding qualitative data covers the judgement calls no tool removes.
Combining both
Switching and combining are not exclusive. You can produce a first coding pass in Themera, export it as CSV or DOCX, and continue in a workbench for memos or visual work. For a longer comparison of the two ways of working, see workbench and pipeline workflows compared, and for how Themera sits next to another established program, the NVivo comparison.
A formulation for your methods section
If software supported part of your coding, say exactly which part. Adapt this to what you actually did:
The material was analysed using [method]. A first draft of the codebook and an initial coding of the transcripts were generated with Themera; each code assignment is linked to a verbatim quote from the material. The researcher revised the codebook, sharpened the code definitions and re-coded the full material. A sample was re-coded independently, and agreement is reported as Cohen's kappa.
In short
Stay with ATLAS.ti if you have a licence, support and a method that relies on memos, networks, queries or non-text material. Try a lighter tool if you have one text-based project, little time and nobody else working in the same file. You do not have to decide on faith: see a worked example analysis with codebook, verbatim quotes and coverage, or try it on your own data: three analyses free, no card needed.