MemFlex

Adds pronunciation audio to every card on the chosen side(s). For languages with a non-Latin script and an auto-converter (Japanese for now), the card text is also enriched with the native script alongside the romanization. Re-running is safe — cards that already have audio or native script are skipped.

Leave off to only voice cards missing audio. Turn on after a bulk text change (e.g. a find/replace across the deck) to rebuild every clip so the audio matches the new text.

Apply one share to every deck you own. Existing shares with the same target get their permission updated.

Current shares

Add a share

AI features run on your own provider account, so you control the model and the cost. Pick a model, then add that provider's API token below.

Each feature below uses that model unless you give it its own.

Your API tokens

Stored encrypted. They're never shown again after saving — only the last 4 characters.

Tap a picture to put it on this card — nothing is saved until you do. All free: 🖼 Pixabay illustrations, ✏️ Openverse & Wikimedia clipart, 🔷 Iconify icons. No match? Generate one instead — that one is slow (up to a minute) but can picture abstract words the libraries can't.