শ্রুতিSRUTIAI

Guided analysis · Research plan

From interviews to findings, one section at a time.

AI drafts codes, themes and claims and shows the passages behind each one. You decide what stays. It all runs on your computer.

Start free trial ↗See the ten sections

The principle

AI suggests. You decide. And we tell you how good it is.

Evidence with every suggestion.

Codes, themes and claims arrive with the passages behind them.

Nothing is final until you say so.

Keep, edit, merge, split or drop anything AI proposes.

Measured, and said plainly.

Where we tested an AI step, the app says how it did. The contradiction finder tells you how many of the pairs it flagged on a test study were real, and how many it missed.

An AI-use statement

for your methods section says which steps used AI.

Choose your approach

Start from the method you’ll write up.

Hybrid thematic

Default

Codes from your brief and literature, plus codes found in the interviews, then merged.

Inductive

Codes and themes built up from the data, for reflexive thematic analysis.

Deductive

Bring your own framework and apply it as written. Import a codebook as REFI-QDA .qdc. For framework analysis and directed content analysis.

Grounded theory is not supported.

The ten sections

Move between them as your study needs.

Work in the order your study needs. AI suggests; you decide.

Brief & literature

Add your brief and papers. AI finds your research questions and drafts a literature review, each finding citing its page.

Brief & literature: two brief documents and four research questions found in them
Participants: attribute suggestions marked Check, and participant cards with city, occupation and employment type
The Codebook: a list of kept codes beside one code with the passages it was coded in
Code passages: a Bangla passage with its English and an accepted AI code suggestion
The Themes section: three themes, each grouping codes with their passage counts, two proposed by AI and edited by the researcher
Claims: a claim with its evidence split into supports, qualifies and contradicts, and the passages under it
Framework matrix: each participant\u2019s account on each theme, with passage timestamps
Export: findings report, codebook, coded passages, full REFI-QDA project, charts and memos, each with its format

Contradiction finder

Find where the story doesn’t hold together.

Disagreement is often where the finding is. SrutiAI reads pairs of passages that share a code and shows you possible tensions, each with both quotes. Read each one before you decide.

  • Two participants who disagree
  • One participant telling two different accounts
  • Sites or groups that differ
  • Interviews against the literature
The contradiction finder: passages from women and men compared on one code, with the test-study accuracy stated above

Your literature, inside the analysis

The papers you already have, and the ones you’re missing.

A folderon your computer, including ones synced by OneDrive, Google Drive, Dropbox or Box

Zotero,read from the Zotero app on your computer

BibTeX or RISwith PDFs, from Mendeley, EndNote, Paperpile or Google Scholar

A link or DOI:SrutiAI fetches the open copy where there is one

OpenAlex search,optional: you see and edit the exact search terms before anything is sent

Charts

See the pattern, then read the passages.

Framework matrix: each participant\u2019s account on each theme, with passage timestamps
Co-occurrence chart: how many participants have passages under both of two codes
Participant \u00d7 code chart: passages coded per participant, read by claim
Evidence balance chart: each claim\u2019s passages split between supports, qualifies, contradicts and unsorted
Treemap: the size of each theme and code by interview passages

Works with your QDA tool

Open your project in NVivo, ATLAS.ti or MAXQDA.

SrutiAI reads and writes the open REFI-QDA standard. Import a codebook as .qdc; export the whole project as .qdpx.

Findings report
Word
Codebook
CSV · .qdc
Coded passages
CSV
Full project
REFI-QDA .qdpx
Charts
PNG · SVG
Memos
Markdown
Export: findings report, codebook, coded passages, full REFI-QDA project, charts and memos, each with its format

Around every section

Memos and a research journal,

attached to codes, claims and passages.

Add recordings later

and code them against your current codebook.

Search across the study

with the optional Search model (334 MB).

Sign in with ChatGPT Not available at launch

An option to run analysis on your own ChatGPT account instead of on device. When you turn it on, the passages it analyses are sent to OpenAI under your account. Off unless you choose it.

Analysis FAQ

Common questions

Does the AI do the analysis for me?

No. It drafts codes, themes, claims and possible contradictions with their evidence. You accept, change or reject each one, and the AI-use statement records what it did.

Which methods are supported?

Hybrid thematic, inductive (including reflexive thematic analysis) and deductive approaches such as framework analysis and directed content analysis. Not grounded theory.

Can I move my project to NVivo, ATLAS.ti or MAXQDA?

Yes: export the full project as REFI-QDA .qdpx.

Can I analyse field notes and documents too?

Yes, beside your interviews. Each AI task reads a kind of source only once we have measured it working on that kind; the app says which.

Which plan includes analysis?

Research.

Try guided analysis on your own interviews.

Included in the Research plan. Payment happens on the website, never inside the app.

Start free trial ↗Compare plans