GPT Transcribe
Speech to text with subtitles and speaker labels, no API key required
Browser-based speech to text tool that adds SRT/VTT subtitles, speaker labels and timestamps to OpenAI's Whisper model, with no API key needed and five free
About this launch
GPT Transcribe is an online transcription workspace running on OpenAI's open-source Whisper model with speaker diarization. It enables users to upload audio or video files, record live, or paste media links to generate timestamped transcripts in over 100 languages. The tool produces transcripts ready to export as SRT, VTT, DOCX, JSON, PDF or plain text, with speaker labels already applied. Every account starts with five free minutes of free transcription, with no installation required and no API key setup needed.
What it does
- SRT and VTT subtitle export
- Speaker separation and labeling
- Record audio directly in browser
- Upload files, record live, or paste media links
- Six export formats: TXT, SRT, VTT, JSON, PDF, DOCX
- Support for over 100 languages
- In-browser transcript editor
- Word-level timestamps
- AI Summary and AI Analytics
- Translation into 100+ languages
Pricing
Pricing model: Freemium
Shopper questions
What is gpt-transcribe?
OpenAI's recommended speech to text model, released July 28, 2026, used for transcribing completed files and Realtime input. It accepts audio plus optional text context and returns text, supports streaming, and takes keyword hints and multiple language hints. It succeeded gpt-4o-transcribe from March 2025.
How much does gpt-transcribe cost?
The gpt-transcribe model costs $0.0045 per minute of audio through the OpenAI API, which is 25% below the $0.006 charged by whisper-1 and gpt-4o-transcribe. However, using it requires building the integration around the API.
Can gpt-transcribe produce SRT subtitles or speaker labels?
No. SRT and VTT output, speaker diarization and word-level timestamps are outside gpt-transcribe's scope and require whisper-1 or gpt-4o-transcribe-diarize. GPT Transcribe closes this gap by running Whisper with diarization, so subtitles and speaker labels come out directly.
Which model does GPT Transcribe run on?
OpenAI Whisper, hosted on Replicate, with a diarization model layered on for speaker separation. It is not the gpt-transcribe endpoint. Whisper was chosen specifically because subtitles, speaker labels and timestamps all come with it.
Is GPT Transcribe affiliated with OpenAI?
No. GPT Transcribe is an independent web tool not built, endorsed or operated by OpenAI. It runs OpenAI's open-source Whisper model with a diarization stage to deliver what people searching for gpt-transcribe usually need: accurate speech to text plus the subtitles, speaker labels and timestamps the model itself does not return.
GPT Transcribe has a free listing. The owner can unlock the full profile with a paid launch for $5:
- Founder story (locked)
- Milestones (locked)
- Demo video (locked)
- Dofollow website link (locked)
Shoppers also viewed
- TokScript
Audio, Voice & Podcasting
- Relayo
Audio, Voice & Podcasting
- hearlog
Audio, Voice & Podcasting
- Audionarrator
Audio, Voice & Podcasting
