Spet

Spet provides free audio-to-text transcription in over 90 languages without requiring users to sign up or create an account. The service is designed to deliver accurate transcriptions, especially for non-English audio, by utilizing OpenAI's Whisper large-v3 model running on a GPU. Spet addresses the limitations of many free transcription tools that primarily focus on English and may perform poorly with other languages, offering strong accuracy for languages such as Ukrainian, Russian, Turkish, Polish, German, Spanish, French, Italian, Portuguese, and Vietnamese.

The platform allows users to upload audio or video files, as well as Telegram voice messages, for transcription. Spet features automatic language detection but also lets users manually select their preferred language if desired. Once the audio is processed, users can either copy the resulting text or download it in TXT or SRT subtitle formats. For Telegram voice messages, Spet can transcribe them directly without the need for prior conversion to MP3, providing a solution for users who find Telegram's built-in transcription limited on free accounts.

Spet emphasizes privacy by deleting user audio immediately after processing. The tool operates entirely online, with no mention of desktop or mobile applications, and is accessible through its web interface. There is no cost to use the service, and no licensing terms are specified in the available information. Spet is positioned as a transcription tool leveraging advanced speech recognition technology for multilingual audio and video content.