Your accent isn’t the problem. Some sound contrasts are.
- Free diagnostic
- No card
- 20 first-language guides
If English isn't your first language, start here
Every first language leaves a few specific, well-documented fingerprints on English pronunciation — not because of effort or talent, but because some English sound contrasts simply don’t exist, or work differently, in the language a learner grew up speaking. Knowing which two or three apply to you is faster than guessing across all 44 English sounds.
What sayit actually checks
A verdict on every word you say, and — when a sound is off — your IPA next to the target IPA. It never marks a word wrong for sounding like it came from India, Lagos, Manila or Warsaw. It marks a word only when the sound that changed is one a listener would hear as a different word.Find your first language
Hindi
For most Hindi speakers, three things do the most damage to clarity: v and w collapsing into one sound, th coming out as a dental t or d, and English's retroflex-free t/d landing too far back in the mouth. Fix those three first.
Urdu
Urdu shares most of its sound system with Hindi, so the biggest three are the same: v and w merging into one sound, th becoming a dental t or d, and retroflex t/d landing in the wrong place in the mouth. Urdu's Persian and Arabic loan consonants add a smaller fourth issue around g and k.
Punjabi
Punjabi's three biggest transfers into English are an inserted vowel before words that start with a consonant cluster (school becomes iskool), the same v-w merger found across the region, and a pitch pattern — left over from tones that replaced older aspirated consonants — that can land stress on the wrong syllable.
Bengali
For Bengali speakers the biggest three are v being replaced outright by b (not merged with w, but overwritten), z shifting toward the j sound, and th coming out as a dental t or d. All three change which word a listener hears, not just how it sounds.
Marathi
Marathi speakers typically carry over three things: the same v-w merger found across much of the Hindi-Urdu belt, a habit of adding extra aspiration (a small puff of air) to consonants that don't need it in English, and th coming out as a dental t or d.
Gujarati
For Gujarati speakers, the three biggest transfers are the region-wide v-w merger, vowels that nasalize more readily near m and n than English vowels do, and a schwa-deletion pattern that differs from Hindi's — Gujarati tends to drop word-final vowels that Hindi keeps, which can make English word endings land unevenly.
Tamil
Tamil's native sound system doesn't distinguish voiced from voiceless stops the way English does (p vs b, t vs d, k vs g), and it has no native f, z, sh or the two th sounds at all — those five gaps do the most damage, ahead of anything else.
Telugu
Telugu syllables strongly prefer to end in a vowel, so English words that end in a consonant often pick up an extra vowel sound at the end; alongside that, f/z/sh are borrowed sounds under-practised relative to English, and voiced/voiceless stops (p/b, t/d, k/g) don't contrast the way English needs them to.
Malayalam
Malayalam speakers carry three main patterns into English: a small vowel added after words that end in a consonant, borrowed sounds like f, z and sh that are less automatic than in English, and the two th sounds, which don't exist in Malayalam at all.
Mandarin
Mandarin's syllables mostly end in a vowel, n, or ng, so other English final consonants get dropped or altered; on top of that, Mandarin has no v (approximated with w) and no th sound at all (approximated with s/z), and its r is a very different, retroflex-like sound from English r.
Cantonese
Cantonese allows more word-final consonants than Mandarin (p, t, k, m, n, ng), but still far fewer than English, so unfamiliar finals and consonant clusters get simplified; alongside that, Cantonese has no v (replaced with w or f) and no th at all.
Japanese
Japanese has a single sound that covers both English l and r, so that pair is the single biggest transfer; close behind are an inserted vowel after consonants that Japanese syllable structure doesn't allow to stand alone, and no native v, which becomes b.
Korean
Korean's l and r are two versions of the same underlying sound (chosen by position in a word, not by meaning), so that pair is the biggest transfer; close behind, Korean has no f or v (both become p/b), and Korean's final consonants are typically unreleased, which can make English word endings sound clipped or swallowed.
Vietnamese
Vietnamese allows only a small set of consonants at the end of a syllable, so English words ending outside that set — or in a cluster like "asked" — often lose their final sound; on top of that, six lexical tones shape stress differently than English does, and Vietnamese has no th sound.
Spanish
Spanish speakers most often transfer three things into English: b and v collapsing into one sound, an extra e added before words that start with s plus a consonant (student becomes estudent), and English's rich vowel system — especially ship vs sheep — being flattened toward Spanish's five vowels.
Portuguese
Portuguese speakers (especially Brazilian Portuguese) most often carry over three things: a short vowel inserted to break up English consonant clusters ("hot dog" gaining an extra syllable), a final l that turns into a w-like glide, and both th sounds being replaced since Portuguese has neither.
French
French speakers most often carry over three things: a uvular (throat-based) r in place of English's tongue-based r, a silent h that leads to dropping English's h entirely, and both th sounds, which don't exist in French and get replaced with s/z or t/d.
German
German speakers most often carry over three things: devoicing consonants at the end of a word (bag can sound like back, rise like rice), pronouncing English w as a German v, and both th sounds, which German doesn't have and replaces with s/z or f/v.
Russian
Russian speakers most often carry over three things: no native th (replaced with s/z, t/d, or f/v depending on the word), no native w (replaced with v), and the same final-consonant devoicing pattern found in German, which can turn bed into bet or rise into rice.
Arabic
Arabic has no p sound at all (b substitutes for it), a smaller set of short vowels than English (which flattens contrasts like pin vs pen), and word-initial consonant clusters are usually broken up with an inserted vowel, since Standard Arabic doesn't start syllables with two consonants together.
Straight answers
Will sayit mark me down for having an accent?
No. sayit scores clarity, not accent — it flags a sound only when the contrast changes what a listener would understand, never for sounding like you're from somewhere. Two accurate ways of saying the same word both score well.
How is this different from a general pronunciation checker?
These guides name the specific sound substitutions that are well documented for your first language, so you can go straight to the two or three contrasts that will move the needle, instead of guessing across all 44 English sounds.
Do all speakers of a language make these mistakes?
No — these are common transfer patterns, not universal ones. Region, education, and how long someone has spoken English all change which of these apply to any one person. Use the diagnostic line to see which ones are actually yours.
My first language isn't listed. What should I do?
Start with the diagnostic line on this page, then open the sound pages for whatever it flags. The 44-sound guide and minimal pairs work the same way regardless of where you started from.