Your AI transcription, hosted in France
A transcription and subtitle suite powered by Whishper, installed and maintained by DINAO. Your audio and video files are processed on our French servers — never sent to a cloud service.
What is Whishper?
Whishper is a fully open-source audio/video transcription and subtitle suite, designed to run 100% locally. It transcribes any media — from an uploaded file or a URL (via yt-dlp) — and exports results in TXT, JSON, VTT, and SRT.
At the heart of the tool, the Faster-Whisper engine (an optimized implementation of OpenAI's Whisper model) ensures speech recognition, while LibreTranslate handles transcription translation, all without cloud dependency. An advanced subtitle editor (highlighting, CPS alerts, segment splitting and insertion) allows you to refine every output.
Architected across several containers (Svelte frontend, Go backend, Python transcription, MongoDB database), Whishper supports CPU and NVIDIA GPU to accelerate the heaviest models. Hosted in France by DINAO, it offers you quality transcription while keeping your media on French territory.
Host Whishper at DINAO
Resource tiers compatible with Whishper prerequisites (minimum 2 vCPU / 4 Go / 10 Go). Hosted in France, fully managed.
- 2 dedicated vCPU
- 4 Go RAM
- 40 GB NVMe
- Daily backups
- Managed & monitored by DINAO
- 4 dedicated vCPU
- 8 Go RAM
- 80 GB NVMe
- Daily backups
- Managed & monitored by DINAO
- 8 dedicated vCPU
- 16 Go RAM
- 160 GB NVMe
- Daily backups
- Managed & monitored by DINAO
1 hidden tier(s) (insufficient resources for this app) : Découverte
This application uses AI
The container hosts the application, not the AI engine (which requires dedicated GPUs) : inference is handled externally, with your own provider key. Prioritize a sovereign engine — Mistral AI, NumSpot, Scaleway, or OVHcloud AI Endpoints (France, GDPR) ; an international provider (OpenAI, Anthropic…) only if a specific capability requires it. Inference subscriptions are not included in hosting.
Technical details
You might be wondering…
Are my audio files sent to an external service?
No. Whishper is designed to run 100% locally: transcription (Whisper) and translation (LibreTranslate) run on your instance, in France. Your media never leaves the server.
What transcription engine is used?
Whishper relies on Faster-Whisper, an optimized implementation of OpenAI's Whisper model. Several model sizes are available, from the lightest (CPU) to the most accurate (GPU).
Do I need a GPU?
Not necessarily. Small models run on CPU for occasional use. For regular volumes or large models, GPU acceleration (Studio/Production plans) is highly recommended.
Where is the data hosted?
On DINAO infrastructure in France, in one of the available datacenters. Your media and transcriptions never leave the territory.
Can I change plans or export my data?
Yes. You can upgrade or downgrade at any time, and your transcriptions remain exportable (TXT, SRT, VTT, JSON) — no proprietary lock-in.
