How does sayit's coach build a drill plan from my results?
Beyond the one headline fix, the coach ranks up to three problem sounds, pairs each with minimal-pair drills, and lists the actual words worth re-practicing.
Underneath the one-fix headline, sayit's coach builds a fuller drill plan from the same take: up to three focus sounds ranked by how often each one actually caused a problem, a matching set of minimal-pair drills for each, and a short list of the actual words from your take that are worth re-practicing. The headline tells you where to start; the plan tells you where to go after that.
30-second version: The coach counts every phoneme-level error across your take, ranks the sounds by frequency, and returns up to three focus sounds — each with a plain-language name, an articulation tip, and up to four contrastive word pairs (like "ship" versus "sheep") that specifically train that sound. It also pulls out the actual words from your take that scored lowest, deduplicated, so you have real material to redo rather than only abstract drills.
How is this different from the single "one fix" headline?
Scope. The headline is deliberately one thing — the single sound tripping you up most — so you're never facing a wall of feedback with no clear starting point. The full plan is what's available once you're ready for more than the headline: a ranked list of up to three sounds (not just the top one), each paired with its own drills, plus concrete words from your actual recording rather than generic examples.
What's in a minimal-pair drill, specifically?
For a focus sound, the plan surfaces contrastive word pairs where that sound is the only thing distinguishing two real words — /ɪ/ versus /iː/ turns into "ship" versus "sheep," for instance. These pairs exist because the sound in question is exactly the kind of contrast that changes meaning in English, so practicing the pair directly trains your ear and your mouth on the distinction that actually matters, rather than practicing the sound in isolation with no contrast to calibrate against.
Where do the "words worth re-practicing" come from?
Directly from your own take — the words that scored lowest among everything flagged, with duplicates removed, capped to a manageable handful rather than dumping every flagged word from a long passage at once. These aren't stock example words chosen to demonstrate a sound; they're words you personally just got wrong, which makes redoing them a genuinely relevant rep rather than a generic drill.
What if my take was mostly clean?
The plan says so honestly rather than manufacturing a drill from nothing — if there's little or nothing to flag, the coach doesn't invent a focus sound just to fill out the plan. A clean take with only a couple of minor words gets a shorter, lighter plan; a genuinely clean take gets encouragement to try something harder instead.
What a full plan contains
| Piece | What it gives you |
|---|---|
| Focus sounds (up to 3) | Ranked by how often each caused an error, with an articulation tip |
| Minimal-pair drills | Up to 4 contrastive word pairs per focus sound |
| Practice words (up to 6) | Your own lowest-scoring words from that take, deduplicated |
Does the plan change every time, or stay fixed once generated?
It's built fresh from each take, so it reflects what actually happened in that specific recording rather than a static profile of your weaknesses. If the same sound keeps showing up as a focus sound across multiple takes, that consistency is itself useful information — a recurring focus sound is a stronger signal than a one-off appearance, and worth a dedicated flashcard deck if it keeps coming back.
Try it
After a scored take in sayit, look past the headline fix to the full plan underneath it — the ranked sounds, the minimal pairs, and the actual words worth trying again.
Hear exactly which sounds to fix.
Say one sentence and get sound-by-sound feedback in seconds. No install, no card.