CurTask vs ATLAS.ti: from coded papers to a written literature review

This page is for students and researchers who are choosing a tool to code and synthesise the papers in a literature review, and who have looked at ATLAS.ti. ATLAS.ti is a well-known qualitative data analysis package, available on Windows, Mac and the web, with AI coding and a wide range of supported data types. CurTask is built around one job: taking a set of papers from reading to a cited, written review, with every AI value traceable to its page.

At a glance

One project from reading to writing

ATLAS.ti is strong at analysis, and many researchers then write in a word processor. CurTask joins those two halves. Each project has three lenses: Library for your papers, Analysis for coding, the matrix and the synthesis map, and Writing for the manuscript itself.

You code a passage in the reader and it appears as a coloured strip beside the text; hand coding is on every plan, including Free. On the paid plans, "Code all papers with AI" codes every paper of the project in one run after showing the credits, each passage "Suggested" until you accept or dismiss it, and papers already coded with the same codebook are skipped at 0 credits. The AI can also sort your codes into categories and themes, each theme backed by quotes from your papers. You group codes into themes on the synthesis map, mark which papers agree, contradict or extend each other, and send a theme to Writing with its passages. In Writing you work in sections you can reorder in binder, corkboard or outline view, cite with @ as you type, and choose from APA, MLA, Chicago, Harvard, IEEE, Nature, AMA or ACS. When it is ready you compile to Word (DOCX), PDF, Markdown or HTML.

Every AI value shows the page and the quote

CurTask's literature matrix lets you describe a column in plain words, such as "sample and context" or "main finding", and run it across your papers. Every value carries a page number and a verbatim quote that was checked to exist on that page. If the AI finds no support, the cell says "No source found". Nothing is filled in on a guess.

AI values stay marked "Suggested" until you accept, edit or restore them. The same rule applies in Writing: AI text is marked until you take it, because the researcher is the author. For a thesis that a supervisor or examiner will read closely, that trail from value to quote to page matters.

Evidence you can follow from the draft

A citation in CurTask is more than a reference: it keeps its page number and its link to the passage you drew it from. The Evidence panel shows the passage for each citation in the paragraph, and "Jump to the page" opens the PDF there, joined by a dashed evidence line. Optional marks show sentences with no citation yet.

The writing sidebar builds on the analysis you have accepted (matrix values, coded passages, themes and links): for each paragraph it shows up to three cards, contradictions first, each saying where it comes from and showing its page and verified quote.

Papers that remember what you learned

In CurTask every paper lives once in your Global Library. Its tags and AI analyses, its "baggage", travel with it into every project you add it to, and reusing an analysis you already have costs 0 credits. Duplicate checks make sure the same paper is never stored or analysed twice. Before any bulk AI job you see the estimated credits; if your balance can't cover it, it doesn't start and nothing is charged.

Built in Malaysia, for research here

CurTask runs in the browser with an interface in English and Bahasa Melayu. Scanned PDFs are read in English and Malay, and the full text is kept when files are compressed. Your papers, notes and writing are stored on servers in Malaysia, backups included, and are never used to train AI. Plans are priced in ringgit: Free with no card, Student at RM25 a month, Researcher at RM49 a month.

When ATLAS.ti might suit you

If your project involves interviews, audio or video, images, survey responses or social media data, or your team analyses primary data together, ATLAS.ti covers that breadth well.

Using both

You can keep primary data in ATLAS.ti and run the literature side of the thesis in CurTask. Drag your papers' PDFs into CurTask (Word, OpenDocument, Markdown and plain text are accepted too); details are filled in from the DOI and duplicates are caught. Then code, build the matrix and write the review chapter with citations in the same project.

Start with the papers for one chapter on the free plan, no card needed.