The sceptical question
Do baby heartbeat apps work
Yes, sometimes, in the third trimester, in a quiet room, with patience. And no, often, for reasons that have nothing to do with you. The useful answer has both halves in it.
Three real recordings, three verdicts, decided live in your browser. One of them is a refusal.
Firstsong is not a medical device, not a doppler and it does not use ultrasound. No app here can tell you that your baby is well, whatever it shows on screen. If you notice fewer movements than usual, call your midwife or maternity unit now, and do not try an app first.
The short version
A modern phone microphone really can pick up fetal heart sounds through an abdominal wall. That is not marketing, it is ordinary acoustics, done in research for decades under the name fetal phonocardiography. The hard part was never whether the sound is there. It is that it is very quiet, your own circulation is very loud, and on a bad day the correct output is nothing.
So the honest answer is sometimes, and the question worth asking of anything you pay for is what it does on the days when the answer is no.
What actually decides whether it works
- How many weeks along you are. The biggest factor by far, which is why every honest claim here comes with a gestational range attached.
- Where your placenta is. An anterior placenta puts centimetres of dense tissue between the microphone and your baby and adds a low whoosh at your own rate. Explained here.
- How quiet the room is. Fetal heart sounds sit in the same low band as ordinary building rumble, and there is a level below which no processing helps.
- Where you put the phone, and for how long. Two centimetres the wrong way is the difference between a clear rhythm and nothing, and most people give up at fifteen seconds when a minute of stillness is the minimum.
The detection rates we measured
The engine was run against public databases on PhysioNet before any interface existed. These are the numbers, split by gestational age, because one average across a whole pregnancy would hide the thing you need to know.
| Gestational age | Recordings | Rhythm found | Error against CTG |
|---|---|---|---|
| 28 to 34 weeks | 20 | 90.0 per cent | 3.52 bpm |
| 34 to 37 weeks | 21 | 85.7 per cent | 2.57 bpm |
| 37 to 42 weeks | 75 | 93.3 per cent | 2.47 bpm |
Two things from that table, pulling in different directions. The answer is yes far more often than the reviews suggest. And even in the best band, about one recording in fifteen had nothing usable in it, on research equipment, in controlled conditions.
The limit that has to sit next to those numbers. Every one of those recordings came from a research sensor on the belly, not an iPhone microphone against real skin. How much survives that move is the measurement we do not have yet. We will not estimate it, and when we have it the figure goes on this site whatever it says.
The problem is not that these apps never work
It is that most cannot say no. Read the one and two star reviews here and the same two sentences come back: it did not work, and I think that was my own heartbeat. Those are one complaint, not two, because an app that cannot report failure has only one way to handle a failed listen. It shows you something anyway.
And the something is usually you. Your pulse is the loudest rhythm in your abdomen, and any resting rate between 55 and 90 doubles neatly into the fetal band, so the wrong answer arrives looking like a right one. That failure has its own page.
To be fair about why: no is a terrible thing to sell. Nothing found gets a one star review that evening, and 142 over static gets a screenshot sent to a group chat.
What the market looks like from outside
We measured the largest site in this category with curl on 23 August 2026. Anybody can repeat it in two minutes.
None of that makes anybody a villain. It says this is a category built to be arrived at and left, rather than answered honestly and remembered.
Do not take our word for any of this
The engine is on this site, running in your browser, on three real recordings. One is a baby, one is a mother, one has nothing in it. Watch the third get refused. It is the fastest way to check whether the paragraphs above are true.
Nothing uploaded, no account, no cookie.
How to test any app in this category yourself
- Put the phone on your forearm, then on a cushion. One has a good adult pulse and no baby, the other has nothing at all. A truthful app reports an adult rhythm and then silence. A dishonest one reports a baby twice.
- Count your own pulse and double it. A number close to exactly twice yours means it found you, and that accounts for most of the surprising successes people report.
- Ask whether you have ever seen it fail. An app with no failure in its vocabulary has successes that mean nothing either.
Firstsong ships that test inside the onboarding, before the paywall, on purpose. We wrote it up here. It costs us the sale if we fail it.
So is it worth it
Depends what you want. If you want reassurance about how your baby is doing, then no, and nothing in this category can give you that. A heart rate is one number at one moment, it cannot rule anything out, and the thing that carries real information is movement. That is why Kicks Count campaigns on this and the NHS guidance does not mention listening devices. If you want to hear it, in the third trimester, and you can live with an evening where the answer is no, then it can be worth it. That is a smaller promise than the category usually makes, and the only one that survives the physics.
One email, the day it opens
Firstsong is not on the App Store yet. Leave an address and you get exactly one message when it is, with the price on the first line. No newsletter, no drip, no partner offers. You can reply to that email and a person reads it.
Stored on our own server, never sold, never shared. Your IP is hashed, not kept. Nothing about your pregnancy is asked for or stored.
Common questions
Is the sound real, or is the app playing me something?
With Firstsong it is your own audio, filtered to the 20 to 200 Hz band where heart sounds live and amplified into your headphones. Nothing is added or synthesised. Be suspicious in general though: a crisp, clean, obviously rhythmic sound is a warning sign.
Why did it work for my friend and not for me?
Most likely one of four things: she is further along, her placenta is at the back, her room was quieter, or she put the phone where the baby's back was.
Do the more expensive apps work better?
Not automatically, and price here has been a poor guide. What is worth paying for is an engine that can refuse, an independent check against your own pulse, and a company willing to publish the parts that look bad.
Can it tell me if something is wrong?
No. It cannot diagnose anything and it cannot rule anything out. A single rate at a single moment is not how fetal wellbeing is assessed, which is why clinics use continuous tracing and a clinical picture around it. The ACOG explanation is a good short read. If you are worried, the answer is a phone call.
Why is there no free trial?
Because a trial that quietly renews into a yearly charge is how this category earns its worst reviews, and the honest test takes one evening rather than three days. The first real listen happens before the paywall. That is the trial.
Keep reading
- Is it my heartbeat or the baby. the failure behind most of the it did not work reviews
- Are baby heartbeat apps safe. the physical answer is easy, the real risk is elsewhere
- Anterior placenta. the most common reason a phone finds nothing
- Week by week, what a phone hears. the same table with the reasoning behind each row
Last checked 2026-08-23