Repository navigation
Expand file tree
/
Copy pathpubliccode.yml
More file actions
155 lines (130 loc) · 5.74 KB
/
Copy pathpubliccode.yml
File metadata and controls
155 lines (130 loc) · 5.74 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
# This repository adheres to the publiccode.yml standard by including this
# metadata file that makes public software easily discoverable.
# More info at https://github.com/italia/publiccode.yml
publiccodeYmlVersion: "0.5"
name: aTrain
url: "https://github.com/aTrainTranscription/aTrain"
landingURL: "https://business-analytics.uni-graz.at/en/research/atrain/"
softwareVersion: "1.5.0"
releaseDate: "2026-01-28"
logo: docs/images/logo.svg
platforms:
- windows
- linux
categories:
- data-collection
- knowledge-management
developmentStatus: stable
softwareType: standalone/desktop
intendedAudience:
countries:
- de
- at
scope:
- research
- science-and-technology
description:
en:
localisedName: aTrain
shortDescription: >-
Offline, GDPR-compliant transcription of speech recordings using
local Whisper machine-learning models.
longDescription: >
aTrain automatically transcribes speech recordings using
state-of-the-art machine-learning models, without uploading any data.
It was developed by researchers at the Center for Data Science in Business and Society
at the University of Graz and tested by researchers at
the Know-Center Graz.
All processing runs completely offline on the user's own device. No
recording or transcription is ever sent to the internet, which helps
researchers meet data-privacy requirements from ethical guidelines and
comply with legal requirements such as the GDPR.
aTrain provides a user-friendly interface to the faster-whisper
implementation of OpenAI's Whisper model, delivering high transcription
quality across 99 languages. An optional speaker-detection mode based on
pyannote.audio assigns each text segment to a speaker. Transcripts export
directly into the common tools for qualitative analysis, ATLAS.ti,
MAXQDA and NVivo, letting researchers play the audio for a text segment
by clicking its timestamp.
aTrain runs on the CPU or, for a significant speed-up, on a CUDA-enabled
NVIDIA GPU. It is distributed through Flathub for Linux and the Microsoft
Store for Windows, and is also installable as a headless command-line
tool for automated transcription pipelines.
features:
- Automatic offline speech-to-text transcription
- Speaker detection and diarization
- Transcription in 99 languages
- GDPR-compliant, fully local processing
- Optional CUDA GPU acceleration
- Export to ATLAS.ti, MAXQDA and NVivo compatible formats
- Command-line interface for headless pipelines
documentation: "https://github.com/aTrainTranscription/aTrain/blob/main/docs/README.md"
screenshots:
- share/screenshots/screenshot_1.png
- share/screenshots/screenshot_2.png
- share/screenshots/screenshot_3.png
- share/screenshots/screenshot_4.png
- share/screenshots/screenshot_5.png
de:
localisedName: aTrain
shortDescription: >-
Offline und DSGVO-konforme Transkription von Sprachaufnahmen mit
lokalen Whisper-Machine-Learning-Modellen.
longDescription: >
aTrain transkribiert Sprachaufnahmen automatisch mit modernen
Machine-Learning-Modellen, ohne dass Daten hochgeladen werden.
Entwickelt wurde die Software von Forschenden am Center for Data Science in Business and Society
der Universität Graz und getestet von Forschenden am
Know-Center Graz.
Die gesamte Verarbeitung läuft vollständig offline auf dem eigenen Gerät.
Weder Aufnahmen noch Transkriptionen werden ins Internet übertragen. Das
hilft Forschenden, Datenschutzanforderungen aus ethischen Richtlinien zu
erfüllen und rechtliche Vorgaben wie die DSGVO einzuhalten.
aTrain bietet eine benutzerfreundliche Oberfläche für die
faster-whisper-Implementierung des Whisper-Modells von OpenAI und liefert
hohe Transkriptionsqualität in 99 Sprachen. Ein optionaler
Sprechererkennungsmodus auf Basis von pyannote.audio ordnet jedem
Textabschnitt eine sprechende Person zu. Die Transkripte lassen sich
direkt in die gängigen Werkzeuge zur qualitativen Analyse importieren,
ATLAS.ti, MAXQDA und NVivo. Per Klick auf einen Zeitstempel wird das
zugehörige Audio abgespielt.
aTrain läuft auf der CPU oder, für eine deutliche Beschleunigung, auf
einer CUDA-fähigen NVIDIA-GPU. Die Software wird über Flathub für Linux
und den Microsoft Store für Windows bereitgestellt und ist zusätzlich als
Kommandozeilen-Werkzeug für automatisierte Transkriptions-Pipelines
installierbar.
features:
- Automatische Offline-Transkription (Sprache zu Text)
- Sprechererkennung und Diarisierung
- Transkription in 99 Sprachen
- DSGVO-konforme, vollständig lokale Verarbeitung
- Optionale CUDA-GPU-Beschleunigung
- Export in ATLAS.ti-, MAXQDA- und NVivo-kompatible Formate
- Kommandozeilen-Schnittstelle für Headless-Pipelines
documentation: "https://github.com/aTrainTranscription/aTrain/blob/main/docs/README.md"
screenshots:
- share/screenshots/screenshot_1.png
- share/screenshots/screenshot_2.png
- share/screenshots/screenshot_3.png
- share/screenshots/screenshot_4.png
- share/screenshots/screenshot_5.png
legal:
license: AGPL-3.0-or-later
mainCopyrightOwner: Jürgen Fleiß
maintenance:
type: community
contacts:
- name: Jürgen Fleiß
email: "juergen.fleiss@uni-graz.at"
affiliation: University of Graz
- name: Armin Haberl
email: "armin.haberl@uni-graz.at"
affiliation: University of Graz
localisation:
localisationReady: false
availableLanguages:
- en
dependsOn:
hardware:
- name: NVIDIA CUDA-capable GPU
optional: true