Skip to main content

Text-to-Speech

Text-to-speech (TTS) turns typed cards into audio without you ever hitting record. It's the fastest way to build a deck, the easiest way to stay consistent across hundreds of cards, and often the only way to study a language you can't yet pronounce yourself.

Pick a source​

Audio Flashcards supports three TTS sources. Choose one in Settings > Text-to-Speech:

SourceBest forNeeds
On DeviceFree, offline study. Quality varies by phone.Nothing — uses voices already on your device.
Amazon PollyWider language and accent coverage at low cost.Your own AWS credentials, internet connection.
OpenAI TTSThe most natural-sounding voices, especially for English.Your own OpenAI API key, internet connection.

Polly and OpenAI use a bring-your-own-credentials model — you supply the keys in the app's settings and pay the provider directly for usage. There's no markup or middleman.

Create a card with TTS​

  1. Open a folder and tap Add Card.
  2. Choose Text for the side you want to generate.
  3. Type the text and pick a voice.
  4. Tap Preview to hear it before saving.
  5. Repeat for the other side and save the card.

Side Correspondence​

Side Correspondence is the setting that saves you from picking voices over and over. It maps a default voice to each side position — voice A for the front of every card, voice B for the back — so new cards inherit the right voices automatically.

Configure it in Settings > Text-to-Speech > Side Correspondence. It's especially useful for language learning, where the question and answer should always use voices in different languages.

You can still override the voice per card when you need to — Side Correspondence only sets the default.

Folder language defaults​

Each folder can carry a language setting. When you create or edit a TTS card inside that folder, the voice list is filtered to match. This keeps the picker short and prevents the wrong language from sneaking in.

Set the language in the folder's settings.

Tricky pronunciations​

TTS engines stumble on acronyms, foreign words, names, and technical terms. Rather than rewording the card, wrap the problem word in SSML — tags like <break>, <say-as>, and <phoneme> are parsed by native TTS engines on both platforms, so no special setup is needed.

Regenerate audio on a card​

Picked the wrong voice or want to tweak the text after the fact? Open the card, tap a TTS clip, and choose Edit. You can change the text, pick a new voice, and save — the audio is regenerated automatically. Recorded clips have an Edit option too, which takes you back to the record screen.

Download more iOS voices​

iOS includes many built-in voices, but the higher-quality ones are optional downloads. Once installed, they appear in Audio Flashcards' on-device voice list.

  1. Open Settings on your iPhone or iPad.
  2. Go to Accessibility > Read & Speak > Voices. On older iOS versions this may appear as Spoken Content.
  3. Choose a language to see its available voices.
  4. Tap the download icon (cloud with arrow) next to any voice you want to add. You can preview voices with the play button. Enhanced voices are larger and may require Wi-Fi.

Personal Voices and Siri Voices are not available for use in Audio Flashcards.

Download more Android voices​

Android's TTS catalog depends on which engine you have installed.

  1. Open Settings on your Android device.
  2. Search for Text-to-speech output (the exact path varies — try Accessibility, System > Languages & input, or General Management depending on your phone).
  3. Tap the gear icon next to your preferred engine (commonly Google or Samsung) to manage installed voices and download new languages.
  4. Set the engine you want as the default, then restart Audio Flashcards so the new voices appear.

If your engine offers a "Network" or "High quality" voice tier, enable it — the difference in quality is usually significant.

When a voice goes missing​

Voices can disappear if you delete them from your system, switch phones, or move a deck between devices. The generated audio still plays fine — it's a saved file, not a live render — but the voice itself is marked Unavailable Voice, so you can't regenerate the clip or create new audio with it until you pick a replacement. Open the card, edit the TTS clip, and choose a new voice.

Next steps​