Skip to content
Lizely

Word Builder

Build six unique English words from the letters S T R E A M using a transparent, auditable challenge list.

Privacy: your files never leave your device. All processing happens locally in your browser.

How to use

  1. 1.Use the S T R E A M tiles or type a word that uses no letter more often than it appears on the rack.
  2. 2.Press Enter or choose Submit word; accepted unique words earn points according to their length.
  3. 3.Find six different words to reach 1,000 points before two rejected or duplicate entries deadlock the run.

About Word Builder

Word Builder gives you one letter rack, S T R E A M, and asks you to find six different accepted words before two rejected entries end the run. Type directly into the field or tap the visible letter tiles, then press Enter or choose Submit word. Accepted words appear below the rack and earn twenty points per letter during the run. When the sixth unique word is accepted, the disclosed completion bonus brings the score to exactly 1,000. Restart resets the rack, found-word list, mistakes, and score.

The game uses a closed 45-word challenge table instead of pretending to contain every English word. That scope is important. A full dictionary would require a much larger maintained dataset, decisions about inflections, names, abbreviations, regional spellings, offensive terms, and word-game conventions. This page makes a narrower promise: it accepts the exact curated list documented by the product and rejects everything else. A familiar word outside the list can therefore be valid English while still being unavailable in this challenge.

The featured completion words include STREAM, MASTER, STEAM, TEAMS, SMART, STARE, TAMER, and TERMS. The wider list adds common three-, four-, and five-letter choices that can be formed without using a letter more times than it appears on the rack. The runtime validates two independent conditions. First, the normalized lower-case word must be present in the curated table. Second, its letter counts must fit the multiset S, T, R, E, A, M. This second check prevents a future table edit from accidentally allowing a word that needs a duplicate tile.

Spelling coverage was checked against the SCOWL English Speller Database and the corpus-derived wordfreq project. Those sources help establish that the featured entries are ordinary English words, but they do not transfer their complete vocabulary into the game. Lizely stores only the small disclosed challenge table, so there is no runtime download, remote dictionary request, or silent change when an external dataset updates. Table size, uniqueness, letter feasibility, and eight featured words are guarded by independent test fixtures.

Input is trimmed and compared without case, so stream, STREAM, and surrounding spaces resolve to the same candidate. A word can score only once per run. Submitting the same accepted word again is rejected because the goal is to find six unique words. Blank text, punctuation, a word outside the table, a word that needs unavailable letters, or a duplicate found word adds one mistake. Two mistakes create a visible deadlock. Completion and deadlock freeze later submissions until Restart, preventing duplicate scoring and making the browser result reproducible.

The letter buttons are optional helpers. Each button appends its character while that tile remains available in the current draft; the text field also supports ordinary typing and editing. The server is not involved in checking an entry. The 45-entry Set, letter-count function, score update, and terminal-state rules all execute in the current tab. No typed word, result, score, or history is uploaded. Best score is stored only in local browser storage when that storage is available.

This is a word-formation game, not a spelling tutor, dictionary, Scrabble authority, vocabulary assessment, or language-proficiency test. It does not define whether a word is appropriate in a particular style, dialect, classroom, tournament, or editorial policy. The two evidence sources also have broader and different scopes than this page. If you need definitions, usage labels, pronunciation, etymology, or an exhaustive word list, consult a current dictionary designed for that purpose.

The interface stays intentionally small: one rack, one input, one action, a found-word list, exact progress, a two-mistake limit, Restart, and the shared double-Escape boss key. It is keyboard friendly, works without an account, uses no new package, and fits a 390-pixel mobile viewport without horizontal scrolling. The fixed rack keeps every local and production playthrough comparable while still allowing many different six-word winning paths.

Methodology & sources

Input is trimmed and lower-cased, checked against a fixed 45-entry Set, and independently checked against the single-use letter multiset STREAM. A unique accepted word adds length times 20 points; the sixth accepted word sets the reproducible completion score to 1,000. A duplicate or rejected entry adds one mistake, two mistakes deadlock, and terminal states ignore later input.

Frequently asked questions

Why was a real English word rejected?
This is a closed 45-word challenge, not a complete dictionary. Only the disclosed curated entries are accepted.
Can I reuse the same word?
No. Each accepted word can score once per run; a duplicate counts as a rejected entry.
Are typed words sent to a server?
No. Normalization, letter counting, list lookup, scoring, and history all run in the current browser tab.
How is the score calculated?
Each accepted word adds twenty points per letter during the run. The sixth unique word applies a disclosed completion bonus for a final score of 1,000.

Related tools

Mini Games guides

View all