Resume landing every interview? Now make your words land too. SpeechPolish closes the loop.
AI writes a structured speech in 60 seconds → record your rehearsal → Whisper transcribes on-device → 5-axis DNA score → pinpoint rewrites for weak spots → re-record and see the delta. The only app that closes the write-rehearse loop.
iOS 17+ · Android 9+ · English & Chinese · Privacy-first
What's live today
Seven capabilities covering every step from “I have a speech due” to stepping up with confidence.
AI speech writing
Pick an occasion (Wedding Toast, Eulogy, Business Pitch, Graduation, and 4 more) → fill in key facts → AI writes a full hook / body / close structured draft in under 60 seconds.
Rehearsal recording + DNA score
Record yourself delivering the speech → Whisper transcribes locally → 5-axis DNA score (pacing wpm, keyword hit, duration deviation, filler frequency, pacing variance) vs. the occasion benchmark.
Pinpoint rewrite
Tap any weak axis → AI flags the exact sentences in your draft → proposes rewrites → accept and re-record → see the score delta. The loop that no other speech app closes.
Occasion DNA library
8 occasion-specific structure benchmarks (hook pattern, target wpm, filler tolerance, duration target). AI writing and scoring both anchor to these — specific guidance, not generic advice.
Teleprompter mode
Full-screen scroll at your pace (80–180 wpm), mirror flip for stand-mounted phones, supports English and Chinese. Practice delivery, not memorization.
Draft history + version compare
Every AI generation, manual edit, and accepted rewrite is versioned automatically. Review, diff, and roll back anytime. All text stored locally — never uploaded as plaintext.
Speaker portfolio
Accumulates your personal pacing median, filler tendencies, and top scores across speeches. AI writing and scoring gradually adapt to you — not just to the occasion benchmark.
Four steps from idea to stage
Write, rehearse, refine, re-record — the real loop.
Pick an occasion, fill in the facts
Choose your speech type, describe your audience, add 3 talking points and a target duration.
AI drafts in 60 seconds
The occasion DNA shapes a full hook / body / close draft. Edit any paragraph or have AI regenerate a section.
Record → DNA score
Hit Record, deliver the speech, stop. Whisper transcribes (only the audio signal leaves your device; the file stays local). 5-axis score report in under 30 seconds.
Rewrite weak spots → re-record → see the delta
Tap any low axis, AI highlights the exact lines and suggests rewrites. Accept, re-record, and the score delta is immediate. Loop until you're confident.
On the way
More occasion types and advanced rehearsal features are rolling out.
More occasion types
v1.0 ships with 8 occasion benchmarks. v1.1 expands to 13: TEDx, classroom teaching, year-end review, parent-teacher meeting, product launch.
Real-time emotion & tone feedback
Live detection of monotone delivery, rushed pacing, and overlong pauses during rehearsal — so you can self-correct mid-take.
Expanded language support
v1.0 supports English and Chinese speeches and UI. v1.1 adds Japanese DNA benchmarks and transcription, followed by Spanish and German.
Want early access? Subscribe to Pro and you're auto-enrolled in new-feature early access.
Simple, honest pricing
Managed by App Store / Google Play. Cancel anytime. 7-day free trial on Pro.
Free
For occasional speakers
- 5 AI speech drafts / month
- 3 rehearsal recordings / month
- 5-axis DNA scoring (every recording)
- Teleprompter mode
- Draft history (last 3 speeches)
Pro Monthly
Best for active speakers
- Unlimited AI speech drafts
- Unlimited rehearsal recordings
- Full rewrite → re-record loop
- Speaker portfolio (long-term)
- Complete draft history
- 7-day free trial
Pro Yearly
Best value for frequent speakers
- Unlimited AI speech drafts
- Unlimited rehearsal recordings
- Full rewrite → re-record loop
- Speaker portfolio (long-term)
- Complete draft history
- 7-day free trial
Frequently asked
1. Do my recordings get uploaded?
No. Your audio files (m4a) stay on your device. Only the audio signal is sent to Whisper for transcription — the file itself never leaves your phone. The transcript also stays local; we do not store your speech text on our servers.
2. Does my speech text get stored in the cloud?
AI drafts require sending your talking points and occasion choice to the AI service to generate the speech. After generation, the draft is stored locally on your device (SQLite). We do not upload your speech text to our servers.
3. What's the difference between Free and Pro?
Free gives you 5 AI drafts and 3 scored rehearsals per month — enough for occasional speeches. Pro unlocks unlimited drafts, unlimited recordings, the full rewrite → re-record loop, speaker portfolio, and complete draft history.