Upload or record
Bring an authorized file or capture the conversation while it happens.
AUDIO TRANSCRIPTION FOR PUBLISHING WORK
Turn audio into a transcript you can publish.
Upload a podcast, interview, meeting or voice recording. CastTranscript keeps the transcript, speakers, subtitles, chapters and show notes in one private editor.
30 minutes free Up to 3 hours Timing and speakers when available Private by default
New transcript Private
Start from audio or video
MP3, M4A, WAV, MP4, MOV or WebM
or drag and drop it anywhere here Private by default Video removed after 24h English + 中文
One working document
Edit the transcript, not a pile of disconnected files.
Correct a name, replay a sentence or relabel a speaker in the document. Playback, subtitle cues and generated notes stay attached to the same edit.
TranscriptSaved just now
The detail worth keeping
NoraWe kept coming back to the same question.
MayaIt is usually the detail you almost edit out. That is the line the listener remembers.
NoraLet’s move it into the opening chapter and keep the pause before it.
From source to deliverable
A clear path through every recording.
Route the job
Batch, live and fallback models stay separate, visible processing passes.
Review in context
Play from a cue, correct the words and rename speakers in one document.
Export the result
Download TXT, Markdown, SRT or VTT without publishing a public page.
MP3M4AWAVMP4MOVTXTMDSRTVTT
Workflows by source
Start with what you have. Finish with what you need.
Each tool changes the input, review surface and expected output. They are not duplicate upload pages with different headlines.
01
Podcast → release notes
Get the transcript, chapters and show notes needed for the episode page.
Podcast transcription 02Interview → usable record
Keep speaker turns, timestamps and quotes attached to the source recording.
Interview transcription 03Meeting → decisions
Turn a finished meeting recording into searchable text and action-ready notes.
Meeting transcription 04MP3 → subtitles
Review sentence timing and export SRT or VTT when the source provides cues.
MP3 to SRT 05YouTube file → transcript
Upload authorized media or import caption text you can already access.
YouTube transcript 06TikTok file → spoken lines
Turn an authorized short video into editable text or captions.
TikTok transcriptCapability-aware routing
Use the right model without building around its name.
CastTranscript chooses a model for the job, records the model used and shows what came back. You keep one editor and one export workflow even as providers improve.
Compare current live modelsBATCHMAI-Transcribe-2 todayFast long-form transcription with word timing and diarization
LIVEGrok Voice Transcribe 2.0 todayStreaming speaker turns while the conversation is happening
SPECIALMuse and other models when the job needs themLanguage, availability and capability determine the route
Audio transcription for real work
One workspace for the recordings people actually need to finish.
Convert MP3, M4A and WAV recordings into editable text, or turn an authorized video into a transcript and subtitle cues. The same private workspace handles podcasts, interviews, meetings and short social clips without forcing every job into one model or one output.
For long media, the service processes queued sections and merges verified timing back into a single document. When a provider returns sentence timing but not word timing or speaker labels, CastTranscript shows the limitation instead of manufacturing alignment with an unrelated model.
When the text is ready, copy it or export TXT, Markdown, SRT or WebVTT. Projects remain private to the account, and exports stay files rather than accidental public transcript pages.
Questions, answered
Transcription questions
What can I transcribe with CastTranscript?
Upload common audio and video formats, record a live conversation or import caption text you are authorized to use. Each source opens as a private, editable project.
Which transcription model does CastTranscript use?
The current default is MAI-Transcribe-2 for uploaded media and Grok Voice Transcribe 2.0 for live sessions. Muse Voice Transcribe remains a configured live fallback. CastTranscript records the model that actually ran.
Can I export a transcript as SRT or VTT?
Yes when the project contains verified cue timing. Untimed caption imports remain available as TXT or Markdown rather than receiving invented timestamps.
Does CastTranscript identify multiple speakers?
Speaker labels appear when the selected model returns diarization data. You can rename them in the editor. Missing speaker data is disclosed without blocking the transcript.
Does uploading create a public transcript page?
No. Projects remain inside the owner’s account. CastTranscript supports copy and file export, not public transcript URLs.
Private by default
Make the next recording useful before it gets buried.
Start with 30 free minutes. Upload a file, record live, correct the transcript and export the result.