MemFlex - Your Personal Memory Trainer

Sign in to MemFlex

Your personal memory trainer

Username

Password

Create an account Forgot password?

Create an account

Email address

Password

Your native language

Language you want to learn

Reset your password

Only works for accounts whose username is an email address.

Email address

Set a new password

New password

Welcome, friend

Your personal memory trainer

Your Decks

Recents only ⭐ Favorites only Show community decks

0 selected Regenerate Parent Deck:

0 cards 0 mastered

Language: ⚠ Overriding your account default

Add language:

Cards

Slide Show Speed: 9-43 sec

Fast Slow

Training Setup

🙌 Hands Free

Test Mode:

Question → Answer Answer → Question Both Directions

Answer → Question isn't offered for this deck's type — a short factual answer (like a date or a number) is too ambiguous to reverse-quiz from.

Advanced Settings

🕰 Adjust weights for time away

Test Speed: 9-43 sec

Fast Slow

Learn Sensitivity: 99%

How much weight is removed on a correct answer

Unlearn Sensitivity: 50%

How much weight is added back on a wrong answer

Target working-set size: 7 cards

How many cards (per direction) stay actively in rotation at once — a finished one is replaced by a new one, so you're never facing the whole deck at once.

Min mic confidence: 45%

Spoken answers below this confidence are rejected as "too noisy" instead of scored. Higher = stricter; 0 = never reject.

Graphic size: 65% of screen

How tall image/hotspot cards can be. Bigger = larger picture; too big can push the buttons below the screen.

Mic debug overlay

Shows live speech-capture diagnostics on the study screen, for chasing a stuck mic. "Copy mic log" copies the full recent diagnostic history (recorded even with the overlay off) for pasting into a bug report. Applies to this device only.

Record UI interactions (usability)

Records what you tap and where you navigate (this device only) so confusing spots can be found. "Usability story" reads your most recent recorded session and describes — in plain language — what, if anything, tripped you up. No answers or card content are recorded, only which controls you used.

🎤 Speaking (spoken answers)

Answering in tests:

⌨️ Type 🎤 Speak, type, then submit 🎤 Speak — auto-submit

Both Speak modes start listening automatically; Type waits for you to tap the mic. The mic stays available in every mode — tap it any time to re-record.

Show pronunciation (accent) feedback

This applies to this deck:

Which side is your native language?

Tells hints which phrase you're learning, and keeps a card's picture off the side you're being tested on.

⚠️ This deck has Answer set as the native side — the standard is Question = your language, Answer = the target. Switch above unless this was deliberate.

Require pronunciation score ≥ (blank = feedback only):

0 / 0 mastered

--s

↻

--/100 pronunciation

Odds of coming up

--%

Odds --

Projected --

Since last --

Adjust --

Shown --

Language

🌳 You did it!

Every card in this deck — mastered. Go stretch those legs.

0 Total Presentations

Add Card

Question & Answer

Question:

Audio: 📁

Multiple Choice Options:

Answer: For multiple choice, enter the correct letter (A, B, C, D, or E)

Audio: 📁

Hints

Image: 👁 📷

Hot-spot detection method: beta Free (slow) — local, no cost No hot-spot detection Quick — fuzzy, low cost Accurate — slow, costs ↳ Free engine — seed model: Outline detail:

Map mode:

Add Hotspots: Auto-detect Manual

Image Edit Mode:

Full Image Hot Spots

0 / 0

Language:

Add language:

Probability (weight controls how often this card appears):

Q→A odds: -- A→Q odds: --

Q→A projected: -- A→Q projected: --

Card Type:

Fixed Display Time (seconds, 0 = auto):

Create Deck

Deck Name:

Deck Type:

✨ Generate with AI

Create a deck of cards from a topic — pick a Deck Type and name above first.

Manage Decks

Share All My Decks

Grant another user or group access to every deck you own at once.

Move Decks to a Folder

Select decks on your deck list, then move them all into one folder at once (or back to the root level).

Export Deck

Pick one of your own decks to download as a JSON file you can back up or import elsewhere.

Import Deck

Paste JSON data or upload a file:

My Groups

Create Group

Manage Users

Add User

Username	Role	Teacher	Can generate	Native language	Target language

›

Add Audio & Script

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.

Question-side language:

Answer-side language:

Regenerate ALL audio (overwrite existing)

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.

✨ Generate Cards

I want to learn:

My language is:

Replace the deck's existing cards (otherwise the new cards are added)

Get Card Content From:

Choose File(s) Detect Hotspots

— OR —

Card Topic Description

More options

How many:

Formality:

Edit Deck

Deck name: Deck Type: Parent deck:

✨ Generate cards

Create new cards from a topic or word list.

🔊 Audio

Add native-script and pronunciation audio to this deck's cards.

Share All My Decks

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

Current shares

No shares yet.

Add a share

Also share all 0 deck(s) under this one

Usability story

Analyzing your most recent session…

Translation missing

This deck hasn't been translated into yet.

Would you like to translate it now?

Ask your teacher to translate this deck.

Decks ready for you

Send feedback

How should we get back to you?

We'll reply to — your account's own email.

Phone

Set your languages

One last step — this only shows once.

Your native language Language you want to learn

Original image

Original card image, before hot-spot labels were erased

🤖 AI Models & Tokens

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.

Use for everything:

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

Spoken-answer recognition:

Currently using: —

Your API tokens

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

API Spend

User:

Model:

From: To:

System tokens

Checked: only spend billed to the shared system tokens (the free allowance) — the budget shared across users. Unchecked: everything, including each user's own token.

✨ Find a picture

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.