A free, open-source tool for qualitative coding, with optional AI assistance that runs on your own computer.
AQDA runs as a local web app in your browser. Your projects stay on your computer, and no cloud subscription is needed. Its AI tools use Ollama on your own machine; they send passages elsewhere only if you deliberately connect AQDA to an Ollama server on another computer.
aqda-explainer.mp4
- Your first project — import documents, code, organize codes, write memos, export. Start here.
- AI tools — setting up Ollama and using the six AI tools.
- Working together — sharing a project through Google Drive, Dropbox, or another shared folder.
- Letting language models suggest codes — advanced and optional; most projects never need it.
- Import text (
.txt), Word (.docx, formatting is dropped) and PDF documents, images (JPG, PNG, GIF, WebP), audio (MP3, WAV, M4A) with optional local transcription, and spreadsheets (.csv,.tsv,.xlsx) with one document per row. Images can be viewed but not coded. - Code by selecting a passage and picking a code, or creating one on the spot. Click a coded passage to see its codes or remove one.
- Build a codebook of hierarchical codes with colors and definitions. Drag and drop to reorder, merge two codes that mean the same, and restore deleted codes.
- Coding fields (optional): answer a few questions each time you apply a code, for example whether the speaker agrees.
- Memos for the project, a document, a code, or a single passage, with
@references to codes and other memos. - Document variables and tags such as author, date, or source, which AQDA can read from filenames on import.
- Several coders: each coding records who made it, and you can show or hide each coder's codings.
- Segments browser: every coded segment across documents, filtered by code or coder.
- Export to REFI-QDA for MAXQDA, ATLAS.ti, and NVivo, and to CSV and JSON.
AQDA uses Ollama to run AI models locally. With Ollama on your own computer (the default), no internet connection is needed and nothing is sent to any server. If you point AQDA at an Ollama server on another machine, the AI features send passages to that machine.
- Topic Search — find passages that match a topic or theme you describe.
- Code Suggest — find uncoded passages that might belong to a code; Apply or Dismiss each.
- Consistency Check — flag coded segments that stand out from the other segments of the same code, as a prompt to re-read them.
- Hierarchy Suggest — get suggestions for grouping your codes into parent categories.
- Define Code — draft a definition from a code's coded passages.
- Summarize Theme — a summary of at most 300 words of what a code's passages say.
The tools question your coding rather than do it for you; the researcher always has the final word. Documents you list under Exclude documents… in the AI panel, such as interview guides, are ignored by every AI feature.
- Python 3.10 or newer
- pipx (installs Python apps in isolated environments)
- Chrome, Firefox, or Brave — Safari has known issues with large file imports and downloads
- Ollama (optional, for AI features) — ollama.com/download
Open a terminal and run:
pipx install aqdaDon't have Python or pipx?
Mac:
# Install Homebrew (skip if you already have it)
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"After Homebrew installs, it prints commands to add it to your PATH — copy and run those lines, then:
brew install python pipx
pipx ensurepathClose and reopen Terminal, then install AQDA.
Windows:
Download Python from python.org — check "Add python.exe to PATH" during installation. Then:
pip install pipx
pipx ensurepathClose and reopen Command Prompt, then install AQDA.
Getting "command not found" after install?
Run pipx ensurepath, then close and reopen your terminal. This adds pipx's install directory to your PATH.
aqdaThis opens your browser at http://127.0.0.1:8765. To stop, click Close AQDA in the app,
or press Ctrl+C once in the terminal. Closing only the browser tab leaves AQDA running, which
is harmless; reopen http://127.0.0.1:8765 to return.
New to AQDA? Your first project walks you through it.
- Download and install Ollama
- Open a terminal and pull the models:
ollama pull nomic-embed-text # for similarity search ollama pull qwen3.5:9b # for analysis and definitions
- In AQDA, go to Settings and select your models under "Embedding Model" and "LLM Model"
- Open the AI panel (sparkle icon in the left sidebar)
More in AI tools.
To transcribe audio files locally using Whisper:
pipx inject aqda "aqda[audio]"Then import an audio file (MP3, WAV, M4A) and click the transcribe button.
pipx upgrade aqdapipx uninstall aqdaUninstalling removes the app but keeps your data in ~/.aqda/. To remove everything, also
delete that folder.
AQDA saves every change as you work; there is no save button. All your projects live in one
private working database, ~/.aqda/aqda.db. You never need to open or move this file.
On Windows, ~/.aqda/ is the folder C:\Users\<name>\.aqda.
Do not put the live aqda.db in Google Drive, Dropbox, OneDrive, or a network folder. To
work with others through such a folder, use AQDA's collaboration feature (below).
- Automatic backups — AQDA keeps seven verified daily backups in
~/.aqda/backups/. It makes an extra backup before updating the database format, before merging codes, and before replacing a project with a collaborator's version (the ten newest collaboration backups are kept; adjustable in Settings). - Restoring a backup — close AQDA, keep the current
aqda.dbas an extra copy, and copy the chosen backup into its place asaqda.db. - Moving or archiving a project — Export → Save a copy to send (.aqda) gives a standalone copy; Import Project on the start page opens it.
- Deleted projects and codes go to a trash and can be restored; codes come back together with their child codes and coded segments.
- Deleting a document is permanent. Its coded segments are removed. Memos linked to it are kept, each with a note naming the deleted document.
Put a folder for the project in Google Drive, Dropbox, OneDrive, your university cloud, or another synced location. In AQDA, open the project, click Collaborate, and choose that folder. Your colleague adds the same folder under Settings → Collaboration and opens the project from Shared projects available. Each of you should set a name in Settings so that codings show who made them.
AQDA works on a local copy and publishes complete snapshots to the shared folder after a short delay; you never save manually. Incoming changes appear automatically, after AQDA has made and verified a full backup.
AQDA does not merge two independently edited versions. If two people work at the same time, or one computer was offline, AQDA keeps both: your project, and a Collaborator reference project to compare it with. You then choose which version everyone continues from. Working together explains the steps.
| Format | Use case |
|---|---|
.aqda |
Save a standalone copy to send or archive — full AQDA round-trip import/export |
.qdpx |
REFI-QDA text exchange — import into MAXQDA, ATLAS.ti, NVivo |
.qdc |
Codebook XML — share code hierarchies between projects |
.csv |
Coded segments as a table — for further analysis in R, Excel, etc. |
Statements .csv |
One row per coded segment with its document variables, coding fields, coder, and model run — for example to build discourse networks in R |
.json |
Analysis data and document variables — for R, Python, or custom processing |
QDPX exports text and audio transcripts as text sources, with document variables and with
memos linked to their codes and passages. Original audio and image media are not included; use
.aqda when an exact AQDA round-trip is needed.
Most projects never need this. AQDA's own AI tools do not code for you. If you want to develop codes inductively with the help of language models, a small program on your computer can let one or several models read your documents and suggest codes with the passages they rest on. This is a shortcut, not a substitute for reading: each suggestion is marked as the model's and sits in the text, where you accept or remove it. The program works with local Ollama models and with online services; with an online service, the text of your documents is sent to that service. See Letting language models suggest codes.
MIT
Built with substantial assistance from Claude Code (Claude Opus 4.6 by Anthropic). Architecture, backend, frontend, and AI integration were developed collaboratively through human-AI pair programming.
Inspired by QualCoder and the qualitative research community's need for modern, accessible, AI-augmented analysis tools.