LumaSpeech is two products. The first is a free articulation word list generator — no signup, 27 sound options, initial, medial and final positions, words, phrases or sentences, copy or print. The second is a paid AI home-practice app for SLPs at $15 per active student per month after a 14-day free trial. We checked every word in the free generator — 822 of them — against a pronunciation dictionary: 809 passed a strict check, one is a clear position error, and the rest are borderline transcription cases. The real limit is size: a single-sound list tops out at 12 words.
Disclosure: we build Ga-loo, which also publishes free word lists and a worksheet generator, so LumaSpeech is a direct competitor. We tested its lists with the same dictionary our own lists are checked against, and every flagged word is listed below so you can judge the calls yourself.
LumaSpeech at a glance
Question | Answer (September 2026) |
|---|---|
What it is | A free word list generator plus a paid AI practice app for SLPs |
Is the generator free? | Yes — “No signup, no limits” |
App price | $15 per active student per month; 14-day free trial for up to 3 students, no card |
Sounds | 24 single-sound options (voiced and voiceless TH separate, plus vocalic R) and R, S and L blends |
Positions and levels | Initial, medial, final; words, phrases or sentences |
List length | Up to 12 words per sound and position (18–20 for blends) |
Export | Copy or print |
How words are checked | Not described on the page; 809 of 822 words passed our dictionary check |
What is LumaSpeech?
LumaSpeech describes itself as an AI speech therapy practice app for SLPs. Students practise articulation or fluency at home between sessions, the app scores each attempt at the phoneme level, and the SLP gets recordings, a weekly caseload digest and progress reports that export to Word. It is aimed at SLPs and SLPAs in private practice, schools and clinics, and the company says self-serve versions for adults and parents are coming.
The word list generator is a free tool on the same site. It works on its own — you do not need the app or an account to use it — and it is the part most people searching “lumaspeech word list generator” are looking for.
Is LumaSpeech free?
The word list generator is free, with no signup and no usage limits. The practice app is not: after a 14-day trial that covers up to three students and needs no credit card, it costs $15 per active student per month. There is no platform fee or contract, adding a student mid-cycle is prorated, and pausing a student stops billing for them while keeping their history.
For a school SLP that math adds up quickly — ten students on home practice is $150 a month — so the app is priced for targeted home programs rather than a whole caseload.
How the LumaSpeech word list generator works
You pick a sound, a position and a level. The sound menu covers R, vocalic R, S, Z, L, SH, CH, J, voiceless TH, voiced TH, K, G, F, V, P, B, T, D, M, N, W, Y, NG and H, plus R, S and L blends. List size can be set to 6, 12, 20 or all words.
The lists are hand-picked rather than generated from a dictionary. The tool ships with 12 words for most sound-and-position combinations — fewer for a few (8 initial Z words, 6 final voiced TH words) and 18–20 for each blend family — so “20” and “All” return the same 12 words for a single sound. Phrases and sentences are built by dropping each word into a short carrier: eight phrase frames such as “a big ___” and eight sentence frames such as “Can you find the ___?”.
The vocabulary is clearly chosen for young children and picture cards. Initial R, for example, is rabbit, red, rain, ring, road, rocket, rope, run, rug, robot, race and radio.
How accurate are LumaSpeech word lists? We tested all 822 words
Method: we took every word published in the generator in September 2026 and looked it up in the CMU Pronouncing Dictionary, the same North American English dictionary behind our own lists. A word passed if the target phoneme sits in the requested position: the first sound for initial, the last sound for final, anywhere in between for medial. For blends we checked the first two sounds.
Result | Words | Which ones |
|---|---|---|
Passed the strict check | 809 of 822 (98.4%) | — |
Clear position error | 1 | peanut on the medial T list — the /t/ is the last sound, so it is a final T word |
Vocalic R on the medial R list | 3 | around and hurry are transcribed with the r-colored vowel of “her”; syrup has two listed pronunciations, one of each kind |
Glide in the spelling, not a separate sound | 8 | flower on medial W; beyond, kayak, royal, loyal, crayon, papaya and coyote on medial Y |
Not in the dictionary | 1 | lunchbox (a compound; the CH in it is genuinely medial) |
That is a strong result for a hand-built list. Only one word is plainly in the wrong list. The other flags are judgment calls, but two of them matter clinically.
Medial R. A child working on consonantal /r/ — the R in carrot or mirror — is working on a different target from the vocalic R in her or hurry. Mixing the two in one list blurs what is being practised. Our guide to teaching the R sound explains why the two are usually treated separately, and we keep them on separate lists: medial R and vocalic R.
Medial Y. Seven of the ten medial Y words carry the “y” inside a vowel glide (royal, crayon, kayak). Many speakers do produce a slight /j/ there, but the dictionary does not treat it as a separate consonant, and for a child who cannot yet produce /j/ those words give a weak model. Canyon and lawyer, also on the list, are the clearer targets.
What LumaSpeech does well
Free, instant and account-free — the fastest way to get a clean home-practice list.
Child-first vocabulary: short, concrete, easy-to-picture nouns.
Voiced and voiceless TH are separate lists (this, mother vs. thumb, bath) — a distinction many free lists, including ours, do not make.
Vocalic R has its own list, separate from prevocalic R.
One click from words to phrases to sentences.
Early sounds (P, B, M, W, H) are covered, not just the late-eight.
Where it falls short
Twelve words is a short list. A session aiming at 100 or more practice trials repeats each word eight times or more — see why trial counts matter in Why Phonetic Validation Matters.
Phrases and sentences reuse the same carrier frames for every word, so they practise the word in a fixed slot rather than in varied speech.
No control over syllable count or competing sounds.
The page does not say how the words were chosen or checked, so a list cannot be audited without doing what we did.
LumaSpeech vs Ga-loo word lists
Feature | LumaSpeech generator | Ga-loo word lists |
|---|---|---|
Price | Free, no signup | Free, no signup |
Words per list | Up to 12 (18–20 for blends) | All matches in our index — e.g. 511 initial R words |
How words are chosen | Hand-picked; method not published | A common-word list checked against the CMU Pronouncing Dictionary |
Vocabulary | Picturable, young-child words | Broader, and includes adult words (regime, roster) you may want to skip |
Voiced vs voiceless TH | Separate lists | One combined TH list |
Syllables and competing sounds | Not shown | Grouped by syllable; competing sounds marked |
Phrases and sentences | Carrier frames | Separate phrase and sentence lists for the main sounds |
Printing | Copy or print | Free printable PDF with trial boxes (e.g. initial R PDF) |
The honest summary: LumaSpeech gives you a shorter, friendlier list; our lists give you a longer, fully checked one that you may need to trim for a young child. Many SLPs will use both.
Who should use LumaSpeech?
Parents and SLPs who need a quick home list: the free generator is hard to beat for speed.
Private-practice SLPs with a home-practice problem: the app's recordings and scoring are the real product, and $15 per student is easier to justify for a handful of clients than for a school caseload.
Anyone who needs long lists, syllable control or an audit trail: use a dictionary-checked source such as our initial S or initial R lists, or compare the options in our roundup of word list generators for speech therapy.
If what you need is a whole printable worksheet rather than a list, Ga-loo builds one by sound, position, age and theme and checks every word before the PDF is made. New accounts get free worksheet credits — no card needed.
Frequently asked questions
Is LumaSpeech free?
The word list generator is free with no signup and no limits. The AI practice app has a 14-day free trial for up to three students, then costs $15 per active student per month.
How much does the LumaSpeech app cost?
$15 USD per active student per month as of September 2026, with no platform fee, tiers or contract. Paused students are not billed.
Are LumaSpeech word lists phonetically checked?
The site does not say. In our check of all 822 words against the CMU Pronouncing Dictionary, 809 had the target sound in the right position, one (peanut) was in the wrong position, and twelve were borderline cases such as vocalic R on the medial R list.
How many words does the LumaSpeech generator give?
Up to 12 words per sound and position for single sounds, and 18–20 for each blend family. A few combinations have fewer, such as 6 final voiced TH words.
Is LumaSpeech HIPAA compliant?
LumaSpeech states “HIPAA & PIPEDA compliant handling” and encrypted recordings. Because the app stores student voice recordings, confirm the data agreement your district or practice requires before enrolling students. The free word list generator collects no student data.
What is a good alternative to LumaSpeech?
For free, longer lists checked against a pronunciation dictionary, see our word lists by sound and position. For more filters and minimal pairs, Articurate Word Lab is a paid option. We compare them in Best Word List Generators for Speech Therapy.
Keep reading
Sources
LumaSpeech — Free Articulation Word List Generator: lumaspeech.com/tools/articulation-word-lists
LumaSpeech — product overview: lumaspeech.com
LumaSpeech — pricing: lumaspeech.com/pricing
Carnegie Mellon University — The CMU Pronouncing Dictionary: speech.cs.cmu.edu/cgi-bin/cmudict
Prices, features and word lists reflect what LumaSpeech published in September 2026 and may change. This review is educational; an SLP chooses the targets for a specific child.