Speaker Copilot
The stage timing copilot for live speakers
Quick launch-day update for everyone who gave feedback here.
Speaker Copilot — a free open-source tool that runs your live talk for
you — is on Product Hunt today: https://www.producthunt.com/products/speaker-copilot
What shipped for launch:
- Auto-paced timed sections: split a talk by percent, change the total
time, everything rescales. On stage, sections advance themselves and a
short vibration says "move on".
- Three card decks one tap away: planned hints, SOS phrases for six
stage emergencies, and Q&A scaffolds for tough questions.
- Rehearsal tracking: rate every test run on five criteria, watch your
average climb toward a readiness target. The live talk gets its own
final score.
- Fully offline PWA, LocalStorage only, no login, $0/month
infrastructure. Two demo talks inside show the whole journey.
Try it: https://speaker-copilot.khomichenko.com
Code: https://github.com/volodymyr-khomichenko/speaker-copilot
Honest feedback — especially critical — is the whole point of today.
If you present, tell me what's missing.
Every speaker knows these three failures:
You glance at the clock mid-talk and realize half your time is gone—and you're still on slide six. Now you're rushing, cutting your best material, ending with "I'll skip this in the interest of time."
Someone asks a sharp question, your mind goes blank, and the silence stretches a beat too long.
The night before a big talk, you realize you've rehearsed the opening five times and the closing zero.
Existing tools don't solve this. Teleprompters make you read and kill live delivery. Timer apps show total time—which tells you you're in trouble without telling you where. Speaker notes describe what to say, not when to move on.
Speaker Copilot does one thing: it runs the talk so you can just speak.
— Pacing. You split the talk into timed sections by percent (change 45 min to 30—everything rescales). On stage, sections advance automatically; a short vibration says, "Move on." No tapping, no reading, no losing track.
— Backup. Three card decks one tap away: planned hints for every stage of the talk, SOS phrases for six-stage emergencies (lost your train of thought, tech failure, running out of time...), and Q&A scaffolds for tough questions—including the honest "I don't know," said well.
— Readiness. Rate every rehearsal on five criteria, jot one note, and watch your average climb toward a readiness target. When the number turns green, you're ready—not because you feel ready, but because ten timed runs say so. The live talk gets its own final score.
Stack, chosen for zero cost forever: React + Vite + Tailwind, LocalStorage only, and PWA with full offline mode (conference Wi-Fi is a myth). No backend, no login, no accounts—your talks never leave your device. Hosting on Vercel's free tier: $0/month infrastructure, permanently.
A note on how it exists: I work in marketing and don't code—AI wrote the implementation across ~40 iterations while I did the product thinking. The repo is public if you're curious what an AI-assisted build looks like from a non-developer.
It's free and open source. Launching on Product Hunt in early September, but the launch is an excuse—what I actually want is feedback from people who present.
Try it (two demo talks inside show the whole journey): https://speaker-copilot.khomichenko.com
Code: https://github.com/volodymyr-khomichenko/speaker-copilot
More information: https://khomichenko.com/speaker-copilot/
If you give talks, pitches, or demos, how do you keep track of time on stage today? That answer directly shapes what I build next.
-
1
The strongest part is that pacing, rehearsal, and recovery are treated as one problem rather than separate tools. The offline-first approach also makes sense for the exact environment where the product is most useful: on stage, when you cannot rely on a connection or another device.
-
1
Thank you — you've articulated the core bet better than I did. "One problem, not separate tools" is exactly it: the moment you're over time IS the moment you go blank, and the rehearsals are where both get trained. Splitting that across a timer app, note cards, and willpower is why the existing setup fails.
Honest detail on offline-first: it started as a trust decision (no backend = nothing to leak), but the on-stage argument won me over completely after testing at a real venue — the phone sat in airplane mode for the whole run and nothing changed. Conference Wi-Fi and roaming are exactly the environment you can't design around, so the app assumes zero connection by default.
Curious since you clearly think in systems: do you present yourself? If yes — is there a fourth piece of the problem I'm missing? (I've been going back and forth on whether audience Q&A capture belongs in the same loop or is a separate job.)
-
1
Would be interested in continuing the conversation beyond the thread. What’s the best email to reach you at?
-
-
Comment