v0.4.13  ·  pre-release  ·  local-first

Qualitative coding that never leaves
your machine.

NAVAL-QDA is a desktop app for coding interviews, documents, and media — build a node tree, run coding & matrix queries, and generate reports, all against a local SQLite database on your own computer.

Platforms: Windows · macOS · Linux License: CC BY-NC-ND 4.0 Storage: local SQLite, opt-in AI only

From raw transcript to reportable finding

Every project moves through the same five stages — the icons below are the same ones on the NAVAL-QDA mark, because this is the whole point of the app.

01

Capture

Bring in what people actually said — interviews, focus groups, open-ended responses.

02

Import

.txt, .docx, .pdf, plus audio & video for transcription — text is indexed automatically.

03

Code

Select a passage, apply a node. Overlapping and non-contiguous coding, tracked to the character.

04

Organize

Build the node tree, merge or move codes, link sources to cases and attributes.

05

Analyze

Query, cross-tabulate, cluster sources by similarity, visualize with a resizable chart dashboard, and check inter-coder reliability before you write it up.

Everything a small qualitative project needs

Hierarchical coding

Build a node tree, apply codes to text selections, rename, merge, or move nodes, and track percent coded per source.

Project Explorer

Every project lives in one place — browse, switch, or create a project without losing your spot in the one you're working on.

Cases & attributes

Classify sources by case and attribute, and review everything at once in the classification sheet and framework matrix.

Analytic memos

Attach notes to a source, a case, or a node — capture the "why" behind a coding decision as you go.

Text & coding queries

Search, word frequency, coding queries, matrix coding queries, and coding comparison with agreement bands.

Interactive dashboard

Bar, pie, treemap, and word cloud charts you can resize and rearrange — click any chart element to filter the rest.

Similarity clustering

Compare sources by word usage or by coding pattern, with a similarity matrix and a dendrogram showing how they cluster.

Optional AI assist

Source summaries and child-code suggestions, pluggable to Gemini or Mistral via your own API key — off by default, and always reviewed before anything is applied.

Local-first by default

Project data lives in a local SQLite database on your machine. Nothing is sent anywhere unless you explicitly turn on the optional AI provider — and even then, only the content you're summarizing goes out.

● no account · no cloud · no telemetry

Download

Grab the source, the full user guide, or a set of ready-made sample transcripts to practice on.

Source code

Clone or fork the full Electron + React project. Build it yourself for Windows, macOS, or Linux.

github.com/navalsingh9/naval-qda View on GitHub

User guide (PDF)

17 pages: full feature reference, five end-to-end worked examples, a comparison with NVivo, and APA 7 citation guidance.

Hosted on SourceForge · HTML Download guide

Sample transcripts

Six ready-made interview transcripts that match the guide's worked examples — import them and follow along.

Hosted on Dropbox · folder Download transcripts

Wondering how this compares to NVivo?

Section 6 of the user guide walks through licensing, platform, coding model, queries, and collaboration — side by side, with no exaggeration in either direction.

Read the comparison