AI models

Change how recordings are transcribed

Help center · For AideMemo 2.0.2

In Settings › Transcription Model, choose AideMemo AI, Apple Speech or a service you added with your own API key, and the language; Regenerate transcribes one recording with another model.

AI models
  1. Which AI model should I use in AideMemo?
  2. Change the model that writes and answers
  3. Change how recordings are transcribed
  4. Use a model from another provider with your own API key
  5. AideMemo says Apple Intelligence isn't available
  6. An online model or transcription service returns an error
  7. Can I keep my memories away from AideMemo AI and other online models?

AideMemo transcribes recordings with the model and language in Settings › Transcription Model; to transcribe one recording with another, use Regenerate… on that recording.

  1. In the Journal tab, tap the gear at the top left to open Settings, then tap Transcription Model.
  2. Choose a model:
    • AideMemo AI: tap Use on its card. It uses credits.
    • Apple Speech, under On This Device: it transcribes on your iPhone or iPad and doesn't use credits.
    • A service under Your API: tap it, or tap Add Model to add one with your own API key (add your API key). The provider receives the recordings you choose to transcribe and may charge for use.
  3. Under Language, choose the language you speak, or Auto where the model can detect it. Each model keeps its own language; Apple Speech lists the languages it offers on your device.

While AideMemo AI is chosen, Label Speakers shows who said each line, using about 2.3 times the credits, and Remember Voices keeps the voices of speakers you name, on your devices and, while Sync with iCloud is on, in your iCloud, to name them in later recordings. Both are off until you turn them on (label speakers).

One recording with another model

  1. Open the recording and tap ⋯ at the top right, then Regenerate….
  2. Under Transcript, choose the Model and the Language.
  3. Tap Transcribe Again, or Transcribe Again & Summarize to write a new summary too.

This changes only that recording; Settings › Transcription Model stays as it is.

What to know

  • Unless you turn off Auto-summarize new recordings in a recording's ⋯ menu, AideMemo transcribes each recording once you stop or import it (how transcription works).
  • The Transcription Model is separate from the Text Model, which writes summaries.

In AideMemo Free, AideMemo AI transcribes while you have credits. Services under Your API show frosted until 5 friends you invited are paying; a choice you saved earlier stays saved, and AideMemo AI transcribes in its place until it unlocks (AideMemo Free).

With no credits left, below 5 paying friends, the 30 minutes you can record each day are still transcribed by AideMemo AI, at no cost, as soon as AideMemo's own server can take them, whatever Transcription Model you chose; Apple Speech and your own models don't run. Recordings from before wait, and are transcribed newest first once you have credits or a plan.

For AideMemo AI, Language lists Chinese in both scripts, Simplified Chinese and Traditional Chinese, next to Cantonese. AideMemo AI writes Mandarin and Cantonese in Simplified characters; with Traditional Chinese chosen, AideMemo turns the transcript into Traditional characters, one character at a time, so the words themselves stay as they were said (软件 becomes 軟件, not 軟體). When AideMemo is in Traditional Chinese, Cantonese and Chinese that Auto hears are written in Traditional characters too; Simplified Chinese, when you choose it, stays Simplified.

This page answers

  • How do I change the transcription model?
  • How do I set the transcription language?
  • Can I transcribe a recording again with a different model?
  • How do I turn on speaker labels?
  • Does Apple Speech use credits?

Something else? See all help articles, or contact us.