How to ask AI questions about a YouTube video
Paste the link into Chat and the AI downloads and transcribes the video. For a long lecture, run Video to Text first and tap Discuss in chat.
Updated November 2, 20263 min read
Say you have an hour-long lecture or an interview in a language you don't speak well, and you want to know what was said about one topic without watching it all. VocaLingo turns the video into text, and the AI chat answers questions about it, makes notes or flashcards and translates the parts you need.
Which way to choose
| Way | Good for | What you get |
|---|---|---|
Paste the link into Chat | Short videos and quick questions | The AI downloads the video, transcribes the speech, looks at frames and answers right away |
Video to Text, then Discuss in chat | Lectures, podcasts, long interviews | A full transcript with the Essence (brief, key points, chapters), and then a chat about it |
Besides YouTube, links from TikTok, Instagram, X, Facebook and other sites work too.
Way 1: paste the link into Chat
Copy the link to the video.
Open the Chat tab and paste the link together with the question, for example: “https://youtu.be/… What does the speaker say about mortgage rates?”
Wait while the statuses “Downloading video” and “Recognizing video” are shown. Then the AI answers.
Ask follow-up questions in the same chat: the transcript stays in the conversation.
A very long video fills the chat's context quickly. If the chat warns that it is almost full, use the second way.
Way 2: Video to Text, then the chat
On the Tools tab open Video to Text.
Tap Paste link and paste the link to the video. The transcription starts by itself, the language is detected automatically.
You can close the app: a notification comes when the text is ready. The Essence with a brief, key points and chapters is created right after the transcript.
Under the result, in the What next? block, tap Discuss in chat. The text goes to the chat, and you can ask questions about it.
Video to Text accepts files up to 3 GB and links. How it works in detail is in how Video to Text works and the Essence of a video.
What to ask
“Summarize the video in 10 points.”
“What did the speaker say about inflation in the second half?”
“List all products mentioned with their prices as a table.”
“Translate the key points into Russian.”
“Make flashcards from the new words in this lesson.”
“Make a 10-question quiz to check that I understood the lecture.”
How much it costs and how long it takes
In Video to Text the transcript costs about 0.73 tokens per minute of video, at least 1 token: a 10-minute video costs about 8 tokens, an hour about 44. The Essence is paid separately and costs a fraction of a token. Questions in the chat are paid like any chat message, by the model and the length of the text, see choosing a model.
The longer the video, the longer the transcription. In Video to Text you can close the app and wait for the notification; if processing takes more than about 20 minutes, the app says it is taking longer than usual.
Frequently asked questions
What if I only need to download the video?
Write “Download this video” with the link, and the chat offers formats. See how to download a video or audio.
The video has no speech, only music and text on screen. Will it work?
In the chat, yes: the AI also looks at frames of the video and can describe what is shown and read the text on screen. Video to Text works only with speech.
Contact support
Describe the problem and we will reply by email. Add the file name or job ID if you have one.