Glossary
What is Video transcript?
Last updated
A video transcript is a text version of the speech in a video, typically broken into timestamped segments that map back to the video. Transcripts are produced either by automatic speech recognition (Whisper, Google Speech-to-Text, AWS Transcribe) or by human transcription, and are the foundation of every other AI feature on a video: search, summary, chapter markers, and ask-the-video chat.
A transcript turns a video into text you can search, copy, and quote. A 90-minute meeting becomes Ctrl+F territory. A tutorial becomes a reference doc. A customer interview becomes a quotable source. Modern AI models like OpenAI's Whisper-1 (the same model Pullsy uses) reach 95%+ word accuracy on clear English speech, 60–90% on accented, noisy, or technical content. The transcript is also what powers video search: type 'quadratic formula' into a transcript search and the player jumps to the moment the speaker said it. The same pipeline (Whisper transcription → LLM summary → chapter detection → ask-the-video chat) is what every modern async-video tool ships with Pullsy included. Related terms: AI video summary, async video, screen recording, tab audio.