Audio and video transcription on steroids.
Knowing what is being talked about. Without having to listen.
Upload audio or video, or paste a link from the web, and in a few minutes you have an accurate transcript with a summary, timestamps and speakers.
Everything you need, built in
98 languages
Automatic language detection and transcription in 98 languages, no setup required.
New: specialized models for Slovak and AmharicAny file, any source
Upload a file or drop in a link. YouTube, Facebook, Rumble and hundreds of other sites are all fair game.
Speaker detection
Multiple voices in one recording? Gaston tells your speakers apart automatically.
Semantic search
Search by meaning, not just exact words, and find matching sentences across your entire library.
AI chat and summaries
Ask questions about any transcript and get answers backed by the exact moment they were said.
Channel monitoring
Track YouTube channels and get notified the moment your keywords are mentioned.
Translation
Translate into dozens of languages, with accurate, natural-sounding results.
Edit, bookmark, organize
Fix mistakes, bookmark key moments, and sort every recording into folders.
Export anywhere
Download transcripts as TXT, SRT or CSV. One file at a time, or a thousand at once.
Accurate transcription in 98 languages
Get a transcript you barely need to touch. On clean speech, the model behind Gaston gets 94 to 97 of every 100 words right in the major European languages, and the language is detected automatically.
-
98.2%of words right on clean English speech (LibriSpeech) -
94.1–97.3%in Italian, Spanish, Portuguese, English, German, French and Dutch (FLEURS)
For languages general models struggle with, we train our own: Slovak, Amharic and Irish Beta.
See all supported languagesSearch your recordings
Find every mention across hours of audio and video. Search for exact words, or describe what was said and semantic search finds it by meaning, even in another language.
-
AND / OR / NOTcombine or exclude terms with Boolean operators -
"exact phrase"match an exact phrase with quotes -
wildcard*use a wildcard to match all word variations
-
paraphrased quotesstill get found, even when nobody said your exact words -
plain languagetype what you mean, no Boolean operators required -
search across languagesone query finds matches in any language you've transcribed, no translation needed
Speaker detection
GASTON automatically identifies speakers in recordings and attributes each sentence to its respective author. Sentences can be filtered by speakers, speakers can be assigned colors, and their names can be edited.
Monitor YouTube channels
GASTON is not just a file transcriber. The monitoring function downloads and transcribes files from selected channels. After you create alerts, you will receive an email notification whenever the words you follow are heard on the monitored channels.
AI for transcript analysis and summaries.
Each recorded file integrates directly with the ChatGPT model, enabling analysis, summarization, and answering questions about the transcription. Submit any transcript-related inquiries using natural language in the language of your choice.
Translation into 43 languages.
All your transcripts can be translated into other languages: most European languages, plus Chinese, Arabic, Japanese, Korean, Hindi and more.
Are you learning a foreign language? Translate the original file and watch the pronunciation of individual sentences in our player.
Organize Your Files, Your Way.
Easily create folders to store and manage your files directly on the platform. Just like on your computer, you can create, move, and organize files within a customizable directory tree - keeping everything structured and right where you need it.
Integrate into your applications
GASTON exposes a REST API so you can bring transcription, search, and monitoring into your own applications and scripts. Official Python and PHP clients are available to get you started in minutes.
from gaston import Gaston
client = Gaston("gapi-...")
media = client.transcribe_url(
"https://youtube.com/watch?v=..."
)
results = client.search("Trump AND Musk")Your recordings stay in the EU
We run our own servers and GPUs in Bratislava, Slovakia. Your audio and video are transcribed, analysed and stored in the European Union, not sent to third-party speech clouds.
Transcription, speaker detection, translation and search all run on our servers in Bratislava.
Your files and transcripts are stored and delivered from data centres in the European Union.
Our only external service is ChatGPT by OpenAI - and only when you use the summary feature and ask the AI about the content of a transcript.
We do not sell, rent or hand over your recordings and transcripts to anyone. The AI assistant above is the only exception, and only when you use it.
Delete a file you uploaded and we erase the recording, its transcript, translations, summaries and bookmarks. Delete your account and everything of yours goes with it. Our backups roll over within 3 days.
One button in your settings, one confirmation, done. No discount traps, no surveys, no calls. Your files stay in your library on the Free plan.
Unlimited and for fair price.
- Unlimited files
- Maximum transcription speed
- Unlimited channel monitoring
- Translation to all supported languages
- Advanced diarization options
- Unlimited AI assistant
- Files export
- Priority support