See the latest updates on Langua

Below you can see the most recent updates. You can also see what's coming next here.

Dec 11th: It has been a frustrating month as 3 of our key AI technology providers have faced major capacity issues due to their rapid growth outpacing their ability to scale their infrastructure. In response to these issues, we now automatically switch the provider if one doesn't respond within a certain amount of time. This doesn't make the system bulletproof and delays can still occur. We're continually monitoring the situation and making adjustments to minimize the impact of these provider issues, and we're exploring ways to make our systems more resilient. Note that you can also switch 'AI model' and 'transcription model' via Settings.

Nov 30th: The first version of the Android & iPhone/iPad apps is live. Please see here for further info and download links (the apps are not yet on the app stores). We are currently working on getting the flashcards live in the 2nd week of December, as well as integrating the purchase system (if you are not a Pro subscriber, please first sign up via the website).

Nov 26th: New work & business roleplays + added History tab to menu (with ability to edit conversation titles).

Nov 25th: Introduced a new tech for transcribing what you say. It's better at picking up exactly what you say, without correcting it. If you prefer the previous tech, you can switch to V2 via Settings. The V2 model is more prone to correct what you've said. Both models may hallucinate (make up stuff if there is background noise), so try to chat in a quiet environment. We've removed the live transcription tech as it wasn't working well.


Nov 20th: Improved conversations, including a couple of new settings:

  • Ask me to repeat corrections
  • Do not always ask me a Q (makes the conversation feel a bit more natural)

These two settings apply to the general chats and debates. They don't work so well in other conversations so we don't apply them everywhere.

If you're a beginner (and have that level in Settings), you'll now see that the AI will say each sentence in your target language before saying it in your native language. We made this change so you can understand without having to constantly hit translate. If you don't want this, switch to Basic level.


Nov 15th: Google has decided to stop allowing sites like ours from importing YouTube videos. So unfortunately, importing videos is currently not possible. 🙁 You can still import texts & search for any podcast. We'll look into other import options in future.


Nov 4th: You can now update tags in bulk on the vocab page, as well as rename tags.


Oct 25th:

  • Improved import of vocab from outside Langua - import directly in Langua (with tags). If translations are missing, we will add them. The same applies to the bulk import from Google Sheets.
  • Export vocab to a CSV in a click (popular with Anki users).

Oct 21st:

  • Custom debates - in addition to the existing debate topics, you can now select ‘custom’ and the AI will give you debate ideas based on topics that interest you.
  • Fixed the tendency for the AI to always default to 'a)' in multiple choice Qs.

Oct 17th:

  • When you can click on a word, then 'all meanings', you'll see an option to save all translations in one flashcard.
  • 10 new languages - Russian, Greek, Turkish, Hindi, Polish, Danish, Norwegian, Finnish, Croatian & Czech. Asian languages coming soon 👀 (but first, the mobile apps...)

Oct 3rd:

  • After clicking on a word, you'll see a new option: phrases. Our AI will show you 3-5 of the most common phrases that include the word you've translated. These can be saved to your account.
  • The 'All meanings' option now uses AI instead of Microsoft. This means it'll work with any combination of languages (Microsoft was only working for translations to English).

Sept 27th: Drag to translate phrases (works only on laptops - for touch devices, use the translator button at the bottom). More info here.


Sept 24th: Download a PDF of the transcript from your conversation. This can be generated together with the feedback report, or as a separate document. The transcript will include any translations, corrections or alternatives you've generated. You can download it via the hang up (phone icon) button near 'record', as well as from below the feedback report.


Sept 15th:

  • More voices/dialects - Spanish, Italian and German now have some new voices. For Spanish, new dialects include Chilean, Peruvian and Uruguayan. We know there is demand for other dialects but it is limited by our tech providers. For Spanish, we've also added a voice model that reduces latency (the time you wait for a response after you reply). You'll see further speed improvements soon. We've kept the older model as the default for now as the newer model is less stable with pronunciation. You can try the new model via the chat settings.
  • Many of you have told us you want more control when using flashcards - to effectively override the spaced repetition algorithm. This is now possible! When you select a filter, this will be applied when you start a flashcards exercise. Words and sentences you create via the 'generate' button will be auto-tagged. You can edit tags via the pencil. We'll add a bulk edit option soon.
  • Search bar added to flashcards page - find vocabulary more easily (useful for editing)

Sept 9th:

  • Instantly generate lists of words (flashcards) related to your specific interests. You can also generate sentences that contain your saved words to help you learn through context. Click the ‘Generate’ button on the flashcards page to give it a try!
  • Improved design of the translation pop up on mobile (further translation improvements coming very soon)
  • Improved corrections and alternative phrasing

August 27th: You can now bulk delete vocab you've saved.


August 22nd:

  • New conversational exercise for intermediate-advanced learners: learn idioms! The AI will give you an idiom within a sentence, and you'll guess its meaning. You can also ask to explain the idiom if desired.
  • Improved formatting of corrections (including bolding corrections, works 90% of the time)
  • A little gamification to motivate you to practice regularly. Earn points by chatting and reviewing flashcards, and check out the leaderboard on this page to see your rank. We also fixed streak tracking and set it up so you will maintain your streak by replying at least 5 times or reviewing 10 flashcards. Previously, streaks were not being counted accurately. Unfortunately, they had to be reset in order to get the new system ready and provide accurate data. Sorry about this. We promise streak data will be accurate and preserved from now on.
  • As you know, we have hands-free options in the chat settings. But if you're recording a long reply on your mobile, your screen might go off. We've made a change to ensure your screen stays on when you're talking.

August 19th:

  • NEW: Pronunciation feedback - currently available for Spanish, French and English (until our tech partner expands into other languages). When you click the lightbulb button for suggestions for what to reply, you'll now see the option to record yourself and get pronunciation feedback before sending the message. It's not perfect, but it is a great way to test and improve your pronunciation!


August 9th:

  • This week, our default AI provider had a major outage. Fortunately, we provide the ability to switch to a model from another provider via the settings (before or during a conversation). You can switch to 'V2' if you ever experience errors, just remember to switch back after as V1 is noticeably better.
  • The AI now knows your name, and its own name. We've excluded this knowledge from roleplays for now.
  • Improved readability - longer responses are now split into separate paragraphs/lines.
  • New explainer video to help you understand how to converse with AI effectively.
  • Fixed generating interactive transcripts for videos. YouTube had made a change that made things difficult for us.

August 2nd:

  • New conversations for improving your grammar! Practice a specific tense or other aspect of grammar via a number of targeted exercises.
  • Improved 'chat about anything, ask me to repeat corrections' chat option. The AI will now get you to repeat the corrected sentence before it attempts to continue to conversation.
  • Added ability to switch AI model via the settings modal. This can be useful if you encounter an AI error during a conversation, for example if the default provider is experiencing downtime.
  • New voice for American English (Morgan). He sounds a bit like Morgan Freeman...

July 30th:

  • NEW: Click the phone icon button at the bottom of the conversation to either close and return to the conversation options, or get a detailed feedback report that includes corrections and how you can improve. You can also download it as a PDF and share it with your tutor. This is only V1 of the report, please do let us know what you think and whether you would like to see anything else included.
  • You now have access to the smartest AI model on the planet. You should notice some differences in the quality and clarity of responses, corrections and explanations.
  • We've added a Preferred Dialect field in Settings. This will allow you to hear words spoken in the correct dialect in flashcards, content imports, and the translation modal. Note, you already have the ability to select your preferred dialect from within the AI chat tool.

What's coming next? 👀

  • Including the article and gender for nouns in the translation modal
  • A faster chat mode, which will feel more like you're having a call with someone.
  • Guided conversation course for levels A1-A2
  • Compelling discussions around your specific interests
  • Adding more languages (Japanese, Mandarin, Korean)
  • In the mobile apps specifically, we're currently building the Flashcards section (due for release very soon) and improving the visual design (we know it's a bit ugly in places). We'll also be:
    • Improving the way suggested replies work so you have to say them.
    • Adding a button to more easily delete the transcript of what you've said in case of errors.
    • Allowing you to listen to what you've said.
    • Adding an audio summary of your conversation.

You can see other things we're planning, as well as suggest and vote on ideas here!