How to use VocaLingo to learn a foreign language

VocaLingo isn't a course with lessons. It helps you work with real speech in the language you're learning: videos, podcasts, voice messages from native speakers, lectures. Below are five common learning tasks and which tool fits each.

How to learn from a YouTube video or podcast

Video to Text gives you the full transcript in the original language and a summary in the language of your app interface. You read the original and check yourself against the summary.

1. Open Video to Text, tap Paste link and paste the link to the video. A podcast file from the device also works.

2. On the Text tab, read the transcript. It's split by speakers, and Show timecodes helps you find the moment in the video.

3. On the Essence tab, read Key points. For lectures and tutorials the summary often adds Terms with the key words and Questions and answers to test yourself.

4. Tap Discuss in chat and ask, for example: "Make 15 flashcards with the most useful new words from this video" or "Make a 5-question quiz on this text". The chat shows interactive flashcards and quizzes right in the conversation.

The transcript costs about 0.73 tokens per minute, so a 10-minute video is about 8 tokens and an hour is about 44; the summary is a fraction of a token on top. More in how Video to Text works and quizzes and flashcards in chat.

How to practice listening with Text to Speech

Text to Speech reads any text up to 10 000 characters in a voice you choose, so you can listen to a dialogue or your own notes as many times as you need.

• For slow, clear speech, choose a MiniMax voice and lower the speed: it goes from 0.5 to 2.

• OpenAI voices take instructions. The Teacher or Calm preset makes the reading even and unhurried.

• To hear the same phrase with a different accent, filter ElevenLabs voices by language and accent.

An OpenAI voice costs about 1.8 tokens per 150 words, at least 2; a MiniMax HD voice costs 12 tokens per 1000 characters. See how Text to Speech works and voice instructions and presets.

How to check your pronunciation

Text Translation has a Dictate button. Say a phrase in the language you're learning. It's a rough check: if the recognized text matches what you meant, your pronunciation is clear enough. Then tap Listen to hear how the phrase sounds in a synthetic voice and compare. Both actions cost fractions of a token. Details: how Text Translation works.

How to practice speaking with the AI

Voice mode in Chat is a spoken conversation with the AI, up to 60 minutes per session. Tap the wave icon next to the message field and say, for example: "Let's talk in Spanish about travel. Speak slowly and correct my mistakes after each answer." You can also ask the chat to remind you to practice: "Remind me every day at 7 pm to practice Spanish". The chat keeps up to 10 active reminders. More in voice mode and memory and reminders.

How to follow a lecture or a tour

Tourist Translator switches to lecture mode by itself when someone speaks longer than 8 seconds, and keeps translating for up to an hour. Leave Show the original on: three lines of the original stay above the translation, so you hear the language, see it written and understand it at the same time. Speech recognition costs about 1.3 tokens per minute, the voiceover is extra. See how Tourist Translator works.

Voice messages from native speakers

If you write to native speakers, send their voice messages to Audio Translation. The Source text tab shows exactly what was said, which helps with fast speech and slang. Record a reply in the What next? block translates your answer into their language, so you can compare it with your own attempt. Step by step: your first audio translation.