Skip to content
CleverKeys Wiki

Next-Word Prediction

Next-word prediction suggests what you are likely to type next — before you press a single letter — from phrases the keyboard has learned from your own typing, plus a small built-in list of common word pairs that fills the gap before it has learned anything. It is fully on-device, opt-in, and off by default.

Quick Summary

WhatDescription
PurposeSuggest the next word from your learned phrases (built-in pairs cover the gap until it has learned some)
AccessSettings > Input Behavior > Word Prediction > Next-Word Prediction
DefaultOff (opt-in)
Data sourceYour own learned phrase patterns first; a small shipped list of common word pairs fills leftover slots. Nothing downloaded, nothing leaves the device

How It Works

As you type, CleverKeys privately learns which words you use after which (word pairs and three-word sequences). With next-word prediction on, those learned phrases fill the suggestion bar at moments when it would otherwise be empty. Learned phrases always come first; if they do not fill all the slots, CleverKeys tops them up from a small built-in list of common word pairs, so the feature is useful on day one instead of after a week of typing:

  • After you finish a word with a space — the bar offers up to 3 likely next words.
  • After you tap a suggestion — the bar refills from the new context, so you can chain whole familiar phrases by tapping suggestions without touching the letter keys.
  • After a swipe — your swipe’s alternate readings stay in the bar (so you can still correct the swipe), and up to 2 learned next words are appended after them.
  • When you tap into text — parking the cursor at the end of what you typed this session can offer continuations, similar to other modern keyboards.

Learned suggestions only appear when the keyboard is reasonably confident: a phrase must have been seen at least twice, and the continuation must be likely (at least a 5% chance given your history). The built-in pairs have no such history to judge, so they are only ever used to fill slots your own data did not — they can never push a learned suggestion aside. An empty bar is still normal — showing nothing beats showing noise. Every suggestion, built-in or learned, is filtered against your dictionary, so a typo you made twice will not haunt the bar.

Walkthrough: typing a sentence

Say you often type “I want to go home”, and next-word prediction is on.

Tap typing:

  1. Type I and press space. The bar shows learned continuations, for example want · am · think.
  2. Tap want. The word commits with a space, and the bar refills: to · a · more.
  3. Tap to, tap go… you can ride your own common phrases tap by tap.
  4. Start typing a letter instead — the next-word suggestions vanish and normal letter-by-letter predictions take over.
  5. Press backspace while no partial word exists — the next-word suggestions dismiss.
  6. End a sentence with . ? or ! — the phrase context resets, so nothing carries across sentences.

Swipe typing:

  1. Swipe “want”. The word inserts automatically, and the bar shows the swipe’s alternate readings (want · went · wart) — tapping one of those replaces the inserted word.
  2. A moment later, up to 2 learned next words are appended after the alternates: want · went · wart · to · more. Tapping to appends “to” after “want” — it does not replace your swipe. Both behaviors live in the same bar; long-press any entry to see which kind it is.

Seeing why a word was suggested: long-press any next-word suggestion. A small sheet shows its source (“Next-word prediction”) and, for a word learned from your own typing, the statistics behind it, e.g. After “want to”: seen 14×, 63%. Before the keyboard has learned anything about a phrase it can fall back to a small built-in list of common continuations; those say After “the”: common continuation (built-in, not learned) instead of quoting statistics they do not have. You can also enable colored origin dots for every suggestion under Settings > Advanced > Suggestion Origin Markers.

When suggestions will NOT appear

ConditionWhy
Feature toggle off (default)Opt-in
”Learn From My Typing” off (Privacy)Next-word reads your learned data — with learning off, that data is not used at all
Password fieldsSecure mode
Private/incognito fields (apps that request no personalized learning, e.g. browser private tabs)The app asked the keyboard not to personalize
Terminal (Termux)Terminal input is not prose
An autocorrect-undo or add-to-dictionary prompt is showingPrompts take priority
Nothing learned yet above the confidence floor, and no built-in pair for the last wordEmpty bar by design. The built-in list covers English, French, German, Italian, Portuguese and Spanish, and only the most common opening words in each

Configuration

SettingDefaultDescription
Next-Word PredictionOffThe feature toggle (Input Behavior > Word Prediction)
Learn From My TypingOnMaster privacy switch (Privacy & Data) — must be on for next-word to work
Context SourceBothWhich phrase model boosts predictions: built-in, your learned patterns, or both
Personalization Strength1.0How strongly your word usage boosts predictions (0 = off, 2 = double)
Suggestion Origin MarkersOffColored dot per suggestion showing which engine produced it (Advanced)

Tips and Tricks

[!TIP] Give it a few days. The built-in pairs cover common openers from the start, but the confidence floor on your own data (a phrase must be seen at least twice) means the suggestions get noticeably more personal the more you type.

[!TIP] You can inspect and delete anything the keyboard has learned under Settings > Input Behavior > Learning & Data — browse learned phrases and words, delete individual entries, or forget everything.

Technical Details

See the Next-Word Prediction Technical Specification.

All docs Report doc issue