Transcribe a YouTube podcast with speaker labels
Turn a YouTube interview into a timestamped transcript. Review speaker changes, correct names and export a checked quote with its source link.
Start with the episode link
Open the episode on YouTube and copy the full video link. Paste it into the transcript reader and choose the original-language captions. If no track is available, sign in and review the duration and allowance before confirming audio transcription.
Separate the voices
Open the speaker tools in the transcript. Listen to a passage where the host asks a question and the guest answers. Label the lines manually, or request speaker detection after reviewing its usage. If you know how many people speak, enter that count when the selected model offers it.
Check the handover
For example, a host may interrupt a guest halfway through one subtitle. Play the timestamp and listen to both sides of the interruption. Correct the affected labels, and leave uncertain lines unnamed until you can identify the voice. An automatic speaker number does not establish a person's name.
Export a verified passage
Select the question and the complete answer, then copy a citation or export Selected Markdown to keep the source and times. Use full SRT when the editor needs timed cues for the whole episode.