There is a specific kind of silence that happens when a speaker realizes they are out of time.
I know it well. Warsaw, a conference room with maybe two hundred people, a talk I had prepared for weeks. Twenty-five minutes in, feeling good, telling a story the room was clearly enjoying. Then I looked at the clock on the back wall.
Twenty-five minutes gone. Twenty left. And I was still on the first third of my material.
What happened next is something anyone who presents will recognize. My speech accelerated. I started skipping slides with that apologetic little phrase every audience has heard a hundred times: “I’ll go through this quickly in the interest of time.” The closing argument, the one thing I actually wanted two hundred people to remember, got compressed into about ninety breathless seconds.
The talk was fine. People clapped. Someone said nice things afterward. But I knew exactly what had happened: I had spent weeks preparing the content and zero minutes preparing the delivery of it in real time.
Why the existing tools don’t fix this
My first instinct was to look for an app. Surely this is solved.
It isn’t.
Teleprompters solve a different problem entirely. They exist so you can read a script while looking at a camera. On a stage, reading is the fastest way to kill the connection with a room. I don’t want to read. I want to talk, and to be told when to move on.
Countdown timers tell you the truth too late. A timer says “18 minutes remaining.” Great. Remaining for what? For the whole talk. It has no idea that I budgeted six minutes for the case study and I’m now nine minutes into it. By the time a plain timer makes you nervous, the damage is already in the past.
Slide notes describe what to say. They say nothing about when to stop saying it.
Paper cards work, sort of, until your hands are busy and your brain is at 90% capacity and flipping through them costs more attention than they give back.
The gap was obvious once I saw it: none of these tools own the pacing. They all leave the hardest cognitive job (am I on schedule right now, at this exact second?) to the person who is simultaneously performing.
The idea: let the talk run itself
The thought that turned into a product was simple. What if the structure of the talk existed inside the phone, and the phone kept the schedule instead of me?
Before the talk, you split it into sections and give each one a percentage of the total: opening 15%, core story 55%, case study 15%, closing 15%. If the organizers cut you from 45 minutes to 30 the night before (this happens constantly), you change one number and every section rescales itself.
On stage, you press start and put the phone down. Sections advance on their own. When one ends, you get a short vibration in your pocket. You finish your sentence and move on. Thirty seconds before a section ends, a gentler nudge warns you. If you glance down, one big number tells you how much of the talk is left, a smaller one tells you where you are inside the current section, and a list shows what’s done and what’s ahead.
That’s the whole core idea. The phone holds the schedule so your head can hold the content.
The second problem: the moments when things go wrong
Once I started using early versions in rehearsal, a second pattern showed up. Timing wasn’t my only failure mode. There were the blank moments.
A question from the audience I hadn’t anticipated. A slide that didn’t load. That awful three-second void where you’ve lost the thread of your own sentence and everyone can see it.
In each of those moments, what you need is not information. It’s a sentence. Something to say out loud while your brain reboots.
So the app grew three card decks, all one tap away:
Clue cards hold planned prompts for each stage of the talk. Your key number for the results section. The one phrase that bridges into the case study.
SOS covers six on-stage emergencies: you lost your train of thought, you need a few seconds to think, you don’t have the information, you’re running out of time, technical failure, a difficult or unexpected situation. Each holds ready phrases. “In the interest of time, let me focus on the key takeaway.” “I don’t want to guess. Let me confirm the details and follow up.” Boring sentences on paper. On stage they buy back your composure in seconds.
Q&A holds answer scaffolds for the questions after: when you know the answer, when you need a moment, when you honestly don’t know, when the question is hostile, when you need to redirect.
The design rule for all three was written in blood: a stressed person should never have to read a paragraph or make a choice. Six situations, one column, huge type. Tap one, get three cards, swipe. That’s it.
The third problem: knowing when you’re ready
The last piece came from a question I couldn’t answer honestly: how do I know I’m actually prepared?
Before Warsaw, I would have said “I rehearsed a few times.” Which is a feeling, not a fact.
So the app tracks rehearsals. You set a goal (ten runs by default). After each timed run, you rate yourself on five criteria and write one line about what to fix. Pronunciation, clarity, timing, confidence, structure. The app keeps a running average and compares it to a readiness target you choose.
The result is a number that doesn’t flatter you. My last preparation went 2.80, 3.20, 3.60, 4.00, 4.40. Watching that climb is the single most reassuring thing I’ve experienced before a talk. And when the live session finally happens, it gets its own separate score, so you can see the gap between how you rehearse and how you perform.
How this thing actually exists
Here’s the part I want to be plain about: I can’t code.
I’m a marketing director. Speaker Copilot exists because AI wrote every line of it while I did the product thinking, across roughly forty iterations of “here’s what a speaker needs at this exact moment on stage” followed by testing the result on my own phone like any other user.
That workflow taught me something I didn’t expect. The bottleneck was never the code. It was knowing what to build, and having the discipline to say no. No accounts. No cloud sync. No AI features inside a product whose entire job is to be predictable when you’re panicking. Every “no” made the tool faster and easier to trust.
Technically it’s deliberately boring: a progressive web app that installs on your home screen, works completely offline (conference Wi-Fi is a myth), keeps everything in your browser’s local storage, and costs zero to run. Your talks never leave your device because there is nowhere for them to go.
It’s free, and it’s yours
Speaker Copilot launches on Product Hunt today. It’s free, open source, and requires no account. Two demo talks are built in, so you can see a full preparation cycle in about two minutes without typing anything.
If you speak anywhere with a time limit (conferences, demo days, webinars, the quarterly review), I would like to know one thing: how do you keep track of time on stage today? Whatever you answer will shape what I build next.
Try it: https://speaker-copilot.khomichenko.com
Code: https://github.com/volodymyr-khomichenko/speaker-copilot
Product Hunt: https://www.producthunt.com/products/speaker-copilot