Benchmarking the accuracy of
AI transcription tools

Select two pieces of software below and see how they compare with each other. (Updated in Jan 2024)

Overall Summary

Best For:

Industry-leading AI and human transcription accuracy for high volume of recordings


Average accuracy across 5 different types of audio


10 hours of transcription

14 day trial


Supported Languages


Best For:

Scalable speech-to-text solutions


Average accuracy across 5 different types of audio


10 hours of transcription

Get $6 in free credit


Supported Languages

1. Audio Type - Standard Interview

Difficulty - Easy

Wrong Text
Corrected Text
Stylistic Differences

98.63% accuracy

Speaker1: Designers feel that the creative process has become quite sterile, like UX has sort of taken away the kind of the creativeness of it. So before, a designer would go to art college
, and everything about that was - the creative process was coming internally. But now everything seems to be validated and everything seems to be, like, very sterile and the way we create things is all about replicating the same user experience patterns and UI elements. And it feels like the validation part has taken away what it means to be creative.

Speaker2: Like, from developer perspective, I must say data is very helpful, so I personally like it. I like to, you know, have hard proof that what I'm working on or the decisions I'm making are having the, the right impact on the user experience, on the app I'm building or on the look and feel.

Speaker2: But I can totally see where you're coming from, when it comes to the creative process. For example, it's quite important and easy to measure stuff that is already out there, to figure out if the decision was right or wrong. But it's very hard to start with data. You know, to start the process of creating something new, from hard facts, you need this ingenuity that you put into the app in the first place. For example, when I create some reference app, I want to showcase some technology or some new way of solving technological problem, but it's a front-end app, . So I need to make it look and feel nice at the same time. And I stay on this blank canvas and then I pray to have some creative designer next to me, you know, coming in with his ingenuity, rather than, just, you know, getting my data and following the data flow.

Speaker1: I mean, it does feel a bit like a chicken and egg situation. What comes first? The creative step of just discovery?

Speaker2: I actually like to think about it as a sandwich, you know, rather than chicken and egg. It's like, you know, you have some bread and then some filling and then some bread again. So I think it all starts with kind of creative process by designer or by some, you know, creative person that is working on a given product
, to have a starting idea, . And then you get data and user testing and all this, you know, heavy weaponry for, polishing the original idea so that it's the best, you can get. But then at the top, you still need this touch of ingenuity. , you know, . Like, when I think about different, visual frameworks on the web, they kind of stop at this second layer. They took the initial idea. They applied a lot of data to make it as nice and, you know, to use and as usable, and as browser-friendly, environment-friendly, whatever you need, and they stopped there. But then the design needs to take such a visual framework and touch this something unique, you know, that - to make a given app or product outstanding. And sometimes I think we stop at this data level and we don't go through this extra step of adding some more art on top of it to make it extraordinary.

Speaker1: Amazing. I mean, I've never really thought of that. It always feels like one comes before the other
. And I know a lot of working environments do feel like a factory in that respect that I draw the picture, then a developer might implement the picture and then I take it back and say, "Could you change the color back?" And - but I think - I suppose it's about finding a process where both can really feel like they're making a sandwich together in a kitchen rather than it being a conveyor belt.

Speaker2: It's a little bit similar to when the agile methodologies entered the software world. , you know, where you go through different iterations of the, of the creative process. And I agree, when there was the situation where first you get a design out and then you give it to developers to implement it and that's it.

Speaker2: It's very stressful for both sides because, developers feel stressed that the design is not really implementable, to be honest, like it's not compatible with the technology, . And then designers are angry because, you know, developers change their design and now it's not as pretty as it was before. So I think it's all about making it more iterative and more collaborative and then it takes the stress out of both parties. And I know there are some - there's some tension on, on the line every now and then, but, to be honest, I really like working with designers. There's this app I worked on in December last year. It's Santa Tracker. It's a, it's a Google app that kind of tracks the steps of the Santa Claus throughout December up to the culmination day where he goes around the world, you know, and gives the presents. And there's a lot of apps you can play, in that experience throughout, whole December. And I was developing this app where a little elf is going through a maze, is navigating a maze, finding presents, and you can look and like the, you see the elf from the top. So it's pretty much just a red dot with a little pointy hat, kind of, and my task was to make it jump through obstacles. And I was totally stuck, you know, because you look from above and it's this is just this red dot that is a hat, how you make it jump? Like, you don't, don't see legs, you don't see arms. I was, I was stuck. I implemented all the technological parts to it in two days, and then for the next two weeks, I was like blank, how to make it look right. And then I asked for designer help, and within a very short time, he came up with the idea that, you know, if you make the dot bigger and they're smaller, it looks like, - you know, it's like floating in the air and then falling down. I was like, "This is so brilliant and I never thought about it." Like, you really need to, trust in your competencies, you know, and ask for help when you need it.

Speaker1 Speaker3: And what I'd love for us to do is to end up in a world where we have language and ideas that are native to the web itself, that embrace the fact that the web is hypertext, not just a another way to build native apps.

2. Audio Type - News

Difficulty - Easy

Wrong Text
Corrected Text
Stylistic Differences

98.17% accuracy

Speaker1: Good morning. Japan has confirmed it will resume commercial whale hunting next July and is withdrawing from the International Whaling Commission. A government spokesman told reporters whaling would be restricted to Japanese territorial waters and its exclusive economic zone. Japan will be joining Iceland and Norway openly defying the international ban on commercial whaling. A warning
, there are images in this report that some of you may find upsetting. Kim Gittleson has the story.

Speaker2: Although Japanese ships hunted whales for decades, the bloody business has officially been banned for nearly 30 years
. with ships like this one only catching and killing whales for what the Japanese authorities claimed were research purposes. But Japan has long wanted to commercially hunt species like the minke whale, which is protected by the International Whaling Commission, though not endangered. The Japanese withdrawal from the International Whaling Commission means that the country will resume commercial whaling by July of 2019. According to a spokesperson, whale hunting will be restricted to Japan's territorial waters and economic zones, meaning ships will cease whaling in the Antarctic Ocean. And Japanese trawlers will only hunt whale species with healthy population numbers. The move brought condemnation from Australia as well as conservationists, with Greenpeace saying that it was out of step and calling on the Japanese government to instead focus on conservation. By withdrawing from the agreement, Japan will officially join Iceland and Norway in resuming a business that has long been controversial. Kim Gittleson, BBC News.

It is Japan will join Iceland and Norway in openly defying the international ban on commercial whaling. We can talk to Mark Simmons, who's is a senior marine scientist with the animal protection group, Humane Society International. Thank you for joining us. What's your reaction to this decision?

Speaker3: Well, I think it's an extraordinary development. We did see forewarning of this at the last whaling commission meeting, which was
, in September when Japan threatened very seriously to leave. But it's still a very surprising and, and very historic development, really.

Speaker1: Well, Japan says that when the agreement was put in place, the decision that the commercial whaling should be stopped was
, was seen as a temporary solution while ways of looking for a sustainable plan were sought. And that just hasn't happened. Is that a fair enough perspective for them to, to say they didn't intend to sign up to a permanent ban?

Speaker3: Well, in
, in, in 1982, when the, ban was agreed, perfectly legal moratorium on commercial whaling was agreed, the countries of the world realized that there was no way that they could manage whaling to make it sustainable. And so certainly there have been efforts since then to go away and try and find a way to introduce sustainability to this. But that's really very, very difficult, thing to achieve. And efforts, for example, to agree any, any management system for whaling have actually failed.

Speaker3: So Japan is stepping away essentially because it's not getting what it wants from the whaling commission. And it makes a whole series of claims about the whaling commission being dysfunctional, which just don't work.

Speaker1: What has the impact of the ban been in terms of protecting stocks? Because when the ban came in in 1986, some species had been driven almost to extinction. How much has that picture changed?

Speaker3: Well, I mean, the good news is that the
, the moratorium can be seen as a big conservation success. So we prefer to call them populations rather than stocks. But the populations of some of the, whales that were hugely depleted by whaling, some of those have increased and that's good. But that doesn't mean that we're at a situation now where we're actually able to simply, click our fingers and bring whaling back into play. And Japan knows that.

Speaker3: There's a lot of spin in the way that Japan is putting out its, its, its statement on its change of heart about the whaling commission. And the whaling commission is the correct international legal body to oversee both whale conservation and whaling. And Japan knows that. The concern is that other countries might follow suit.

Speaker1: What do you think the impact will be specifically of Japan? I mean, obviously, as you say, other countries may follow suit, but specifically Japan saying it will fish in its own waters. It
, it won't, hunt in Antarctic waters in the Southern Hemisphere, but it will, do commercial whaling in its own territorial waters and economic zones. How, how many whales potentially could that impact?

Speaker3: Right. Well,
that's a, it's a complicated message, isn't it? So we can say that we're very pleased to hear that Japan is going to leave the Southern Ocean. There have been calls on them to leave this very special area of sea around Antarctica, for many, many years. So that's really, you know, quite good news as long as they, as long as they stay away.

3. Audio Type - Medical Jargons

Difficulty - Medium

Wrong Text
Corrected Text
Stylistic Differences

95.69% accuracy

Speaker1: Many people think of oxytocin as the love hormone. It's responsible for the natural high a mother gets from breastfeeding. It's also the feeling of connection with a loved one. And now scientists are discovering the potential of oxytocin to treat addiction and social disorders.

Speaker2: Oxytocin is
, what's called a neuropeptide and it's a type of chemical that's produced naturally in the brains of all mammals, including humans. And we know that oxytocin dampens down stress and anxiety, and it also enhances people's motivation to engage in social behaviour behavior.

Speaker1: Oxytocin is a hormone produced in the part of the brain called the hypothalamus. It's then released throughout the brain and into the bloodstream. It links social interactions to the brain's reward system, promoting bonding and affecting our mood.

Speaker2: Oxytocin that's released into the bloodstream plays a role in functions such as stimulating uterine contractions during childbirth. It controls the milk
let letdown down reflex during breastfeeding. But what we're really interested in is the oxytocin that's released in a number of brain regions that play a critical role in motivational behaviour behavior, in social behaviour behavior, and in stress and anxiety responses.

Speaker1: Worldwide, around 35 million people suffer from opioid, cocaine, cannabis and amphetamine
-type drug -use disorders, but only one in seven receives treatment. These addictions can cause death, chronic medical conditions and cost society billions of dollars in health care. But neuroscientist and psychopharmacologist Associate Professor Michael Bowen's research shows that oxytocin may be able to help moderate addictive behaviours behaviors.

Speaker2: We know that some people can have differences in their brain oxytocin systems and these differences can make people more prone to certain types of disorders
. For instance, it can make them more susceptible to developing, addictions in their adulthood.

Speaker1: Over the past decade, Associate Professor Bowen and his colleagues have developed a compound with similar effects to oxytocin and have conducted numerous preclinical trials on rodents. The team's results show that the compound has had anti-addictive effects in animal models of alcohol and methamphetamine use
. And early evidence suggests it may be effective for other substances such as opioids.

Speaker2: We've also had our results independently replicated across multiple different labs and all the signs point to this having great potential.

Speaker1: Addiction has a powerful influence over the brain and the way it registers pleasure. Alcohol is one of the most harmful drugs worldwide with more than
100 a hundred million people suffering from alcoholism. It's also responsible for around 3 three million deaths a year.

Speaker3: We know that drugs and alcohol produce euphoria and trigger surges of dopamine in the brain. So these surges of dopamine can then lead to cravings and teach the brain to want more or seek out more of the substance
, and this can then lead to addiction for some people.

Speaker1: Professor Adam
Guestela Guastella and Dr. Kelsey Kelsie Bolton Boulton are also involved in oxytocin research. They're currently treating more than 100 a hundred people with alcohol addiction with an oxytocin nasal spray for eight weeks to see if it reduces cravings.

Speaker4: Now, oxytocin provides a great target because it can potentially reduce the cravings and anxiety
, and it can do it in a way which doesn't require people to stop taking alcohol. They can take the oxytocin with the alcohol, and, gradually reduce their addiction to the alcohol over time.

Speaker1: It sounds like a magic cure, but Associate Professor Bowen warns there are still some issues that need to be overcome to unlock the potential of oxytocin.

Speaker2: It can't be taken orally because it just gets broken down in the gut
. And even when it's administered through other means, it has a very short half-life, which means it can't act for very long.

Speaker1: Associate Professor Bowen and his colleagues have been working over the past decade to overcome these hurdles. Together with the company he co-founded,
Canoxis Kinoxis Therapeutics, they're now seeking approval from the US Food and Drug Administration to go into human trials in 2021.

Speaker2: When you look across the spectrum of disorders of the brain and mind
, we see this profound dysregulation of social behaviour behavior from addiction to autism to dementia to depression. We're hopeful that this new drug that we're developing will provide a desperately needed major breakthrough in the way that we treat these disorders.

4. Audio Type - Accent

Difficulty - Hard

Wrong Text
Corrected Text
Stylistic Differences

87.58% accuracy

Speaker1: I have to be here for work
, . This is, you know, the basement.

Speaker2: Yeah
. Yeah, your work's lovely. I saw your wig room.

Oh, yes.

Speaker2: I was in your wig closet.

Speaker1: Yeah, yeah, there's an entire... , I used to put wigs on all the time, and then I had to start working here. So, the Boardwalk Empire is a huge hit.

Speaker2: It is.

Speaker1: Well done.

Speaker2: Yeah, yeah.

Speaker1: Where do you make that?

Speaker2: We
, we film in New York, which is great. Not Atlantic City. It's set in Atlantic City, but...

Speaker1: But you don't film in Atlantic City?

Speaker2: No.

Speaker1: You ever been there?

Speaker2: Yeah, we went
, last year. They had a season premiere in New York and Atlantic City, so that was...

Speaker1: That's HBO, they've got the money.

Speaker2: Yeah
, .

Speaker1: Yeah.

Speaker2: But not this year, apparently. We didn't go to Atlantic City this year.

Oh, you didn't?

Speaker2: No.

Speaker1: Well, you see, the thing about Atlantic City as well is it's kind of like Vegas, but by the... , have you been to Vegas?

Speaker3 Speaker2: Yeah.

Speaker1: Oh, what did you think?

Speaker2: Fun.

Speaker2 Speaker1: Really?

Speaker2: Really fun, yeah. They had a fake beach in the hotel I was in. And waves.

Speaker3 Speaker1: Everything's fake in Vegas. Everything's fake in Vegas.

Speaker1: And so that's what they do, we do fake.

Speaker2: Yeah.

Speaker1: It's lovely. I, I like it. I prefer, though, the, the kind of coastal vibe of Atlantic City, though.

Speaker2: Yeah.

Speaker1: I don't know, I've just decided that right now.

Speaker3 Speaker2: Alright, okay.

Speaker1: So how's Glasgow, then?

Speaker1: Is everything all right? Do you live in a giant house or under a bridge?

Speaker3 Speaker2: A giant castle.

Speaker2: Yeah, no, a castle.

Speaker1: Really, a castle?

Speaker2: Well, it's a big house, bigger than we have in New York.

Speaker1: You have
, you've two houses?

Speaker2: Well, no, we sort of own our house in

Speaker1: Now, who is we? You and

Speaker2: My husband.

Speaker1: I see.

Speaker2: Yeah, Douglas. He's also

Speaker1: You married Douglas from Glasgow?

Speaker3 Speaker2: Yeah, do you know him?

Speaker1: Does he play in a band?

Speaker2: He does.

Speaker3 Speaker1: What band?

Speaker2: Travis.

Speaker1: Yes, I do know
him ! I, I don't know him, I just knew that.

Speaker2: Right, okay, good .

Speaker1: I got, I got it from Wikipedia. Good band, though, Travis.

Speaker2: They're a very good band.
They're a good band.

Speaker1: Did you go and
... , is that how you met him? You were... ,

Speaker3 Speaker2: A fan?

Speaker1: Yeah.

Speaker2: I
... , how did I meet him? I met him... , it was a series of bumping in intos twos kind of thing. Like, in the airport, I met him once.

Speaker1: Is that

Speaker2: And then

Speaker1: Is that how you first met him, in an airport?

Speaker2: Yeah, in an airport. They had flown in from Ireland and I was flying off somewhere.

Speaker1: And you just said,
"Oh, hi."

Speaker3 Speaker2: "Hi."

Speaker2: Yeah, I spoke more to Fran, Dougie just sort of was a walk-by.

Speaker3 Speaker1: Who's...? ?

Speaker2: Dougie, Douglas.

Speaker1: Oh, right.

I'm And Fran, singer.

Fran. Oh, you spoke to the singer?

Speaker2: Yeah.

Speaker1: But your husband's the bass player?

Speaker2: Yeah.

Speaker1: Well
... .

Speaker2: Didn't work out, the singer, what can you do?

Speaker3 Speaker1: Bass player is still good.

Speaker2: No, the bass player is...

At least it wasn't the drummer. Yeah. Are you a musical person yourself?

Speaker2: I'm really not. I really

Speaker1: Really?

Speaker2: Really not, really. My son is, though, . My son's very musical.

Speaker1: What age is your son?

Speaker2: Three
and a half.

Speaker1: I see.

Speaker2: Yeah.

Speaker1: Is he playing the bass?

Speaker2: He's playing the pretend tennis racket bass
, you know.

Speaker1: That's
, that's how you know. That's if they, if they do that, then they're going to be bass players. And if they do that, they're not.

Speaker2: Yeah, no, well, I think he's a guitarist.

Speaker1: He's a guitarist?

Speaker2: Yeah, he uses my dressing gown belt as a strap.

Speaker1: That may be something else, nothing to do with being a musician. I mean, you can be a musician and do the thing with your mum's dressing gown. But it's, you know, it could be a different singer.

Speaker2: Maybe.

Speaker1: Yeah.

Speaker2: Yeah, no, actually, yeah, you're right.

Speaker1: I've got two kids. You, one kid or two?

Speaker2: One.

Speaker1: You're going to have more?

Speaker2: I think
, I think he's just sleeping now to seven o'clock, and we're delighted. And to throw another one into the mix would be a bit...

Speaker1: Yeah.

Speaker2: The 4
. :30 wake-up, song so.

Speaker1: And I got a new one right now.

Speaker2: Have you?

Speaker1: Yeah, eight months old.

Speaker2: Oh, lovely. Boy or girl?

Speaker1: Boy. Very unreasonable.

Speaker2: Yeah, they are, they are. You can't talk to them.

Speaker3 Speaker1: Just very, and then, like, what? We did everything.

Speaker1: We, we got rid of the, kind of, the thing, and then we changed it all around and dried them up, and...

Speaker2: I know.

Speaker1: Nothing. Waa!

Speaker2: Yeah, I'm done, I think.

Speaker1: Really?

Speaker2: One and done.

Speaker1: No, I think you'll end up going again.

Speaker2: Possibly.
That's what happens.

Speaker1: Especially if you're in Glasgow, because the winter's coming, and these nights are long.

Yeah. It's true.

Speaker1: It is true. You're in the
... house and it's dark at five o'clock, you think, "Time for another baby."

Speaker2: Yeah. Yeah, yeah.

Speaker1: It's nice, though, Glasgow . Where do you, what do you go out and, uh, do you go out and drink and eat and stuff in Glasgow? And you're in the west end is what I'm seeing.

Speaker3 Speaker2: Are We're in the west end.

Speaker1: Do you ever go down
to at the Ubiquitous Chip?

Speaker2: Yeah, I used to work there.

Speaker1: So did I.

Speaker2: Really?

Speaker3 Speaker1: Yes.

Speaker2: No.

Speaker1: I did, I worked in the bar, the upstairs bar at the chip.

Speaker2: Yeah, I was downstairs.

Speaker3: I was, I was...

Speaker1: That's where the good-looking people went. I worked in the, I was in the upstairs with a

Speaker2: Yeah, yeah, yeah.

Speaker1: I used to be the charge hand of that bar.

Speaker2: Wow.

Speaker1: Yes, I was, uh, I, the Fustenbergers Fürstenberg is, I ended up in rehab.

Speaker3 Speaker2: Oh, God.

Speaker1: Yeah. So you were downstairs, were you a waitress?

Speaker2: I was a waitress, yeah

Speaker1: Oh, nice. Yeah.

Speaker2: Yeah, yeah, yeah.

Speaker1: Did you work with Susie?

Speaker2: I think I might have
done , is she's blonde. ?

Speaker1: Yeah, I think
, you know, you know .

Speaker2: Lemon juice and sunshine, I think, you know. what I mean?

Speaker2: I think, I think, actually, the Ubiquitous Chip has still got my last wages in their safe. I mean

I mean... You didn't collect your last wages?

Speaker2: I
, I left in a bit of a, um...

Speaker1: Did you get fired?

Speaker2: I didn't get fired, but it was just bad feelings all around,
I'm and kind of...

Speaker3 Speaker1: That's how I left. I

5. Audio Type - Multi Speaker

Difficulty - Hard

Wrong Text
Corrected Text
Stylistic Differences

93.84% accuracy

Speaker1: I've been listening and there's a lot of conversation about behavior change and
, I can't help but wonder about the ethics around that choice. Even the phrase make people happy sort of smacks of a certain amount of control, and I was wondering how, you think about the users', right to choose their experience and whether the fact that manipulating emotions, might have some ethical question, especially when you're looking to do it on the employers employer's benefit as opposed to the end users' were or even as the end users.

Thank you, an important question.

Speaker3: So this is actually kind of similar to, to, to where gamification can backfire. , right, ? There are absolutely ethical questions and, and it comes down to the, the fact that if you're not helping somebody along the path that they want to go, if you don't understand their goals and their, their hopes and their desires and, and, and, if you're not driving that, then there's no reason for them to use your product, there's no reason for them to come back. And they, and, and any kind of nudges are going to seem just completely off base, right? Like the blogs that use gamification and you're and, and, you know, try and use it as a one -size -fits -all tool for driving behavior. And you get there and it's like, "Wow, I have a badge and a profile on this blog and, you know, I'm, I'm really becoming a better person."

Speaker1: 70%. 70%, so close.

Speaker3: Right. Almost. But absolutely, if you're, if you're nudging people to, to do something for the wrong reason or that actually isn't good for them, then, you know that's. , that's certainly an ethical issue, that, if your users don't call you out on it, then hopefully you stop before that.

Speaker3 Speaker4: It I think that's an excellent question that you pose as well, and it also brings up another question that I have. , is that is: is it paramount that the things that we make make people happy, or is it paramount that the things that we designed so well should be invisible, that allow people to give back to their lives and, and forging their own happiness? Should these products and services and platforms that we make- , I mean, if we're talking about users' happiness, it sounds like we're trying to make us happy, in which that's? ,

Speaker2: that's not.

Speaker3: You know, those are two different things, so I, I, I have questions about that too. If I design something, if I do my job correctly, you really shouldn't know that anything has happened. You should just be like: , "Wow, like, I have 15 extra minutes in my day," or, "I, you know, got to see my kids more. I, I, I ran an extra mile, ." I'm, it, just, I. , I have the same questions you do, so I'm, we're unpacking it slowly together.

Speaker4 Speaker5: There's one catch to that, though, too. No design decision is benign. It either thwarts or facilitates well-being somewhere down the road, and again, depending on where you believe happiness or well-being goes- , medical model, hedonic model, eudaimonic model- , it all comes down to well-being at the end of the day, . Everything we interact with, all the digital technologies, all our interpersonal relationships that we have- , they're all dialogues- . They all affect our well-being in one way or another. So I would pose the question of: , "Are the? , if the product and service that we are designing does not contribute to the well-being of the individuals, groups or populations that are interacting with it, does it deserve to exist?" That's a question I take to clients all the time. now, .

Speaker2: Who decides? What's the line between care and control?

Speaker5: Autonomy support. If you design a product that someone is going to interact with, that takes that the person in the, the bottom position, right, the person that's interacting with it. If it takes their perspective, if it allows for choice in their inputs, if they can interact with it when, where, how they want that their goals are framed as their goals, that the actions that they take towards achieving those goals, and if they choose to not interact, that needs to be coded as okay, . right,

Speaker2: Right.

Speaker5: Right? So autonomy support to me is the most important thing.

Overall Comparision

85.69%YesEnglish only