Whisper Transcription评价分析

Eu comprei a versão completa no meu macOS, sendo o mesmo aplicativo da mesma empresa não me dá acesso a versão do iOS. Se eu soubesse não teria comprado no Mac. PESSIMO!

一直显示错误，不值，不如电脑版好用

autant la version macOS est une pépite autant la version iOS est une catastrophe sans nom : Le meilleur moyen pour cracher son iPhone. Il a même réussi à me faire planter l'App Store lors des tests vocaux dans la recherche. Bravo! La clé n'a pas l'air de fonctionner correctement et indique qu'un quota dépassé alors que ce n'est pas vrai et ça Swipe de fenêtre en quittant celle utilisée à chaque fois que l'on clique sur le micro. Et donc il faut revenir à chaque fois sur la fenêtres pour faire la dictée vocale. Insupportable... je ne comprends pas vous ne tester pas vos applications avant de les mettre on line ligne

Auto year Renew at background

Doesn’t seem like the real deal better download from the website

I spent hours to convert speech to text and at the end I see it has no understand any part correctly and it is full of errors

Not Live Translate.

there is an error when i try to download this app.

You can download the app. You can even get the app to create a transcript of your audio file. You can see the audio file transcript waiting for you in the queue... but you can’t have it unless you pay. This was not clear anywhere.

I am a long time user and always love using this app to run whisper transcription on it. However in recent updates I tried to batch transcript as usual using whisper v3 large but it keeps crashing (but whisper v3 large turbo doesn’t), plus, whenever I tried to change the name tag of speakers or do literally everything on the UI it lags a lot and even sometimes crash. I tried serval ways to fix including close every app or restart but nothing changed. Hope it get fix sooner.

I wanted a recorder that could transcribe and summarize meetings, all I got was technical chores, confusion and a customer service that started ignoring me. The software recordes audio fine, but will do a terrible job at transcribing. At a certain point it starts hallucinating and repeating the same words. So no reading back or summary at all. It basically just turns into an expensive voice recorder. There are much better alternatives out there that are free. I would’nt have written this review if only customer service didn’t start ignoring me. I want my money back.

App doesn’t work on the several devices I’ve tried, and the developer’s documentation—accessed from inside the app—is vague about its limitations. Error messages just say “couldn’t process the file”, with zero indication of why (device strength, file format, audio encoding, file size, etc.). It’s possible this is great under the hood, but when the hood is krazyglued shut it’s pretty hard to tell.

Have spent time to install. Only to know that it has no trial. I even don’t know whether it works or not and I have to pay first! pathetic

The segments feature, which used to work, has been removed, so I can no longer navigate to and listen to specific parts of my recordings. In addition, I've sent two inquiry emails, but I haven't received any response at all. It has become completely unuseable. I feel very disappointed and regret subscribing to the paid version.

I’ve bought the pro license but it simply doesn’t deliver. Speaker recognition is way off and Dutch language is still way off. Also when redefining the number of speakers, it assigns the wrong speaker to the most parts of the transcripts. Hence, I submitted requests for support, and before the refund period ended, also requested a refund, since the dev never replied. Giving it 1 star because this simply feels like a scam due to the false promises.

Youtube transcription stops working, Even videos that have been successfully transcripted now is not able to be transcripted.

The transcription works great, the user identification is reasonable. The new NVIDIA models are way faster than any of the Whisper variants. The batch mode just plain doesn’t work for me - it crashes every time - resulting in a do you want to send debug logs to apple every time. I have tried to work with the support team - they have just asked me different ways to clean my config and use the default settings - all with the same result - crashing.

Dopo aver scaricato l’app, viene richiesto subito l’abbonamento, ma su iMac il pagamento non va a buon fine. L’unico modo per acquistare l’abbonamento è tramite App Store mobile, ma anche in questo caso, l’app su iMac non riconosce l’abbonamento già attivo. Ho cercato supporto, ma nessuna risposta. Lo sviluppatore non risponde neanche tramite i canali social indicati sull’App Store. Un’esperienza davvero deludente, soprattutto considerando che l’app è la stessa e non esiste alcuna versione separata tra Mac e iPhone.

模型无法下载，一直请求超时，无法使用

I can’t believe that I spent around 1 hour just to install and see how things works. First of all, YOU CANT USE THIS APP. I have installed - recorded my voice for around 1 minute - then started transcription - waited for few seconds - suddenly there is a popup asking to upgrade to a paid plan to see the voice transcription - or we will have to switch the model - first of all, how can we PAY without testing anything? - there is no trail or anything - then I thought I will try to switch to free models - so I downloaded the V3 model (it doesn’t have a star icon just next to it - so its free) and activated it - recorded again - tried to transcribe - TADA - money maker popup again - asking to upgrade into a paid plan - I changed the model again and ensured that I am using V3 - but still they need a subscription even for a free model. IN SUMMARY - THIS APP IS USELESS. YOU CANNOT USE IT.

Non funziona con lingue diverse da quelle più comuni europee e non lavora bene sul’ascolto. Perde parole, perde concetti, e non trascrive quello che viene detto lasciano spazi e trascrivendo solo alcuni pezzi dell’intero audio importato.

I cannot find parakeet model in Version 12.13， my Mac is 15.2

1. On first launch I was prompted to upgrade. Without ever using the app how can I even make that decision? 2. The simplest action is to start a new recording; the app asks for save location before a recording can even begin (instead of storing temporarily and persisting later; but ok, no big deal). Then during recording there isn’t any live transcription, so I have no feedback on how much it has heard. 3. Once you finish recording, the transcription screen asks to upgrade to access the transcription or switch to a free model. Note that I still haven’t been able to see any output whatsoever. 4. In the model selection dropdown there aren’t any free models listed. Only ‘Large V3 Turbo’ is available and selected; but since it’s a ‘premium’ model, you can’t see the transcription. Thanks for nothing.

mac上的会员居然在iPhone上不能用？？？

I had Whisper subscription on my MacBook and I love it. I installed the IPhone App, tried to restore my subscription, but it didn’t work. It does not make sense have a subscription for MacBook and other to IPhone.

The app is so raw that I don't even have a history 😱

Transcripts are useless, at least for the free model versions. Cannot be recommended.

I purchased the MacOS version of Lifetime. I was wondering if this would allow me to use the iOS version as well, but I cannot use it at all. Do I need to pay extra?

I upgraded to a paid version (MacWhisper, which I think now was a mistake) as the unpaid version, recording a meeting (not directly, but from the computer sound) came out partially in Chinese. However, the paid for version is not much better, as the transcription comes out as a solid lump, no paragraphs, little punctuation and I as the personal pronoun comes out in lower case. It takes hours to edit and so is not worth it. The help facility does not seem to work. It is better if you can drop a file on it. Won’t be renewing my subs.

下载了还不能用，还要下载模型，下了模型还不能用，用的时候巨慢，4分钟的音频，显示30分钟才能转完。刚付了费，后悔，还没人家免费的好用。

wants to charge me to use my own hardware

useless. It fails to translate any of my video. How can i ask for refund?

A very good app. Have purchased the app on a Mac, yet cannot download it from a different Mac.

If you want to convert the content in the video into text and then transfer it to other AI tools, etc., it will be quite inconvenient.

worked ok at first, stopped working, tried to reach out to support, discovered there was no support.

I thought it was a useful app, so I purchased the lifetime plan. I wanna use like record a one-hour meeting, transcribed it with this app, and summarized it using ChatGPT. However, when I transcribed the one-hour m4a file using LargeV3 or V2, it started out well, but from the middle, it just kept outputting the same sentence repeatedly. It's unusable.

due stelle di incoraggiamento, ma è assurdo che la versione a pagamento dia risultati peggiori di quella gratuita!!!

I just launched this app for two sec. It started to ask me for comments which is undoubtedly extremely bad.

I started using this app about 6 months ago. It was great because I got pretty good transcriptions and I could export them in paragraphs. The developer started making changes and nothing has really worked since. Paragraphs don't work, speakers don't work and there doesn't seem to be an easy way to contact the developer (I don't and won't use X). When I finish this, I'm deleting it off all my machines because it doesn't work.

After i paid, i can't find the gpt api option.

The most recent update to this software which has severely restricted the ways to export transcript text has taken this software from an incredibly useful tool in my arsenal to a nearly useless one. When the ability to export as "Speaker Paragraphs" without defining speakers returns, I will consider revising this review. Or maybe I won't—because the developers of this software seem to enjoy removing or revising important features on their whim.

first chinese audio transcript to simplified chinese characters, after updating, only traditional chinese character

The main use I had for this app was to record app audio. I got the lifetime paid version for it, but it's not working. When I try to record a Zoom meeting, for example, it records only the other side. If I try to record just me speaking into the mic, it records only a sentence here and there. I checked the settings and mic selection and it's all correct. These developers also don't have a website to contact support, so I'm left with no other option but to express my disappointment via this review.

I've tried various things, but I can't download the language model, and I can't get a response to that error.

I was excited to download this one. The online reviews are amazing but obviously paid for because this program is no good. I transcribed three interviews, large portions of text were repeated for at least fiften lines when the voice section spoke normally. And then the program just quit itself. Never able to open it again - tried reinstalling it. Like it collasped on itself.

Only M4A files can be read. When I try to load MP3 WAV MP4, I get an error "Could not convert audio. Thisformat is currently notsupported. I thought if I paid for the software, it would be able to read the files, so I paid for it, but I am still having trouble with the error message. M4Aファイルしか読み込めない。 WAV MP4を読み込もうとするとエラー「Could not convert audio. Thisformat is currently notsupported.」と出る。有料にすれば、読み込めるのかと思って、課金したが、エラーのままで困っている。 【追記】 Some files were unreadable ven in MA4. The type of file does not seem to be the problem. Please let me know how to get rid of the error. MA4でも読めないファイルがありました。ファイルの種類が問題ではないようです。どうすればエラーが起きなくなるか教えてください。 2023/02/11

Pro membership: Speaker identification has been super inaccurate and frequently missing audio in transcription lately. Please, go fix it!

I have a license bought by my company, I see no option to enter the key. Could you please help me? In the Settings > Pro I see only options to subscribe and to purchase, but not to enter my key. What should I do?

Easy to use, and the accuracy for Brazilian Portuguese was very good (not excellent, but very good). It would be interesting to add a feature to define timestamp lengths… I would like to define a timestamp every 5 or 10 min, so it is easy to find specific parts on the audio. Currently the timestamps are based on dialogs, but sometimes there are some lengthy dialogs. Overall, I will recommend this to friends.

Système intuitif et rapide pour transcrire des audio. Génial!

需要中文

中英文综合用Large V3就挺好

Somewhere between 4 stars and 5. Overall, very impressed. it does most of what I need it to do, and has already saved me hours of time. I got a one month subscription to check it out, and i’ll upgrade now to annual. i’m transcribing a series of talks and q+a. The app does a great job of transcribing the talks 90%+ accurately. it also identifies different questioners’ voices, separates them out, and recognises when it’s the same voice asking a later question (speaker 2, speaker 3 etc…) . it’s all the more impressive because the audio isn’t great: voice memo on an ipad recording a room with 30 people. If i need to check a bit of the transcription, clicking on the word takes me direct to that bit of the audio, and plays it, making checking obscure bits of the transcription (like unfamiliar technical terms) very easy. it also transcribes the audio in a fraction of the recording time, just a few minutes to transcribe a 50 minute multi-voice session. the free version of the app has a lot of functionality, and probably enough on its own for a single voice. the pro app identifies multiple speakers. i’m being a bit mean giving it just four stars: it’s making a big job much, much easier, though there are still a few (very few) inaccuracies, that are pretty easy to find and check. one other little niggle is that it imports with american english spelling, and i can’t see a way to switch to uk english.

thank you for bringing it into iPhone

충분한 가치가 있네요. 제가 사용하던 윈도우 위스퍼보다 빠르고 출력 품질도 좋습니다.

Ik heb net 2 video’s voor een medische cursus in het Nederlands laten ondertitelen. Werkt zeer goed. Video van 17 minuten met slechts 3 handmatige aanpassingen. Ik ga nu de vertaling naar het Engels uitproberen in combinatie met een Chatgpt promo om de stijl van de presetratrice en medische termen te borgen. Wordt dus nog vervolgt maar tot nu toe ben ik erg positief.

This app brings the power of the Whisper AI model to the Mac desktop. The transcription accuracy is great, handling technical jargon, multiple speakers, and challenging audio quality with surprising precision. The experience is perfectly tailored for macOS, feeling like a native utility. Drag-and-drop functionality for importing media is seamless, and the transcription speed (even when using larger models) is impressively fast on Apple Silicon. The ability to process files locally on my Mac without sending sensitive audio to the cloud is a huge advantage for privacy and security. Although my workflow has a few niche requirements that the current set doesn't completely address, the version on the App Store is more than adequate for most people’s needs. The developer is very active in the community and is constantly making improvements to the app. If you deal with audio or video files and need a solid transcription on your desktop, look no further.

I’m using it usefully. thanks. one suggestion. I wish there is another choose option for colored speaker name in pdf. (even it colored in app, but not pdf)

I do wish the keyboard name can be shorter though, Whisper Dictation — Whisper is kind of unnecessarily long for no good reason besides making the menu digustingly wide when switching keyboards.

I have been using Whisper for years, it has been grate HOWEVER!!!!!!! What happened!! This last update Not so good. Speaker editing issue: I was frustrated at first I could no longer edit the speakers on a Whisper format saved file like before they just all show up as “unknown.” After some trial and error, I discovered that if I re‑transcribe the audio file, I can rename the speakers again in the new transcript and they no longer show up as just “unknown.” Still, this feels like an unnecessary extra step. Uses up time and resources unnecessarily. BIGGEST disappointment – transcript + AI separation: In previous versions, I loved having the transcription and AI together on the same page. Now, that option is gone. The only workaround I’ve found is to open the transcript in one tab, go to the AI page, then right‑click the sidebar to reopen the transcript. It works, but it’s clunky and far from intuitive. I wish there was an option to make it more intuitive.

There is no traditional chinese output recently

Va sempre premesso che in un mese l’AI fa passi da gigante, con relativi miglioramenti che potrebbero portare altre app ad avere risultati migliori. Ma ad oggi, dopo un anno di utilizzo per ogni tipo di necessità da quelle lavorative a quelle di studio non ho il minimo dubbio che è la migliore app per Mac. Una velocità ottima che è destinata ad aumentare in base alle capacità del Mac e di quale M si utilizzi. Integra servizi AI esterni, sempre comodi per riuscire a trovare ed aggiustare il lavoro.

希望能增加中文支持和国内方式付款，本人愿意付费 但确实英文不怎么样

Ich benutze die Software nun lange und habe Jordi schon oft angeschrieben und ihm Verbesserungsvorschläge gemacht. Er ist sehr nett und setzt vieles sofort um. Selten habe ich so eine gute Kombination von Software und Support gefunden.

The best on-device whisper transcription app. Especially the keyboard, really well designed. Consider subscribing to use cloud transcription models to save device battery.

为什么看不到模型的信息，是 whisper 么，希望支持 Nvidia Parakeet

I use this app every day. Jordi (the dev) is very responsive and he continues to tweak and improve it regularly. There isn’t anything else on the market that compares in terms of quality and if you add the substitutions in the mix and have a semi-decent recording you’ll end up with a transcript, if not 100% is completely understandable and few mins to get all the way.

I purchased the pro version just a couple of weeks ago for my iPhone and overall been happy with the dictation using the additional keyboard, much better than Apple‘s built-in solution. I wanted to use it on my iPad Pro as well as my iPhone but don’t see any way to be able to activate the pro version on there. When I try to restore subscription, nothing happens which is rather disappointing due to the high price of the pro version. The developer website doesn’t really exist and there doesn’t seem an easy way to ask what a solution.

Quite accurate. It even can transcribe lyrics of Japanese songs on YouTube. One model that I want but not yet included in current version: Belle-whisper-large-v3-zh It would be great if it can be added in next update.

I'm grateful for this app that allows offline, on-device transcription, which is both better for security and better for the environment than cloud-based apps. I do find transcription quite accurate on the word level, though not identifying speakers--frequently identifies many more speakers than there are and missatributes speach. At the same time, I spend a lot of time cleaning transcriptions up--mostly re-identifying speakers and combining or breaking segments up for coherence. This is necessary for someone who needs transcriptions to be readable for others. Other than improving speaker identification, there are several small improvements that could GREATLY speed up this process-- - have keyboard shortcuts to go rewind playback 15ish seconds - link audio to words rather than segments, so that when you break a segment into two, the second part does not still have the start time of the first. This is pretty tedious to correct. - in segment view, where you can edit, it is sometimes hard to play from a selected section, taking multiple clicks to get out of editing or renaming speaker mode - a comprehensive guide to editing transcriptions. I couldn't figure out how to split or combine segements until I wrote to the developer I am hopeful that this app will be improved to the point that it will replace my need to use others when I'm short on time. If it can, I will recommend this app widely--including to friends, students, and colleagues with similar needs. I will say that when I contacted the developer, I received a quick personal response, which was much appreciated. I have paid for a year of pro-- at the end of the year, I will reassess.

Purchased this software from the website, and it's awesome! However, I prefer downloading and updating software from the App Store. I'm wondering if a license purchased from the official website can activate a version downloaded from the App Store? Hope the developer can answer!

Das neue Update mit der Tastatur unter dem Globus erhöht die Nutzbarkeit der Whisper-App deutlich. Die sehr gute Spracherkennung kann man so zwar etwas ungewohnt über den Globus, aber somit in allen Apps nutzen. Ist das Diktat sehr lang, kommt es vor, dass der Bildschirm gesperrt wird und leider ist die dazugehörige Tastatur nur auf Englisch (qwerty) ohne Rechtschreibkontrolle und mit nicht abschaltbaren Haptik-Feedback. Aber das ist trotzdem ein sehr guter Anfang. Thank you :)

I’ve been enjoying this app for a bit and wanted to leave a review. One thing that i think needs more clarification/development is the new transcription keyboard-embedding. Despite the claim that the developer won’t take any data, i am worried: What if somehow something goes wrong (in ios or the app). It would leave the information i type open for abuse (since you have to give full permission to the app). Perhaps a restricted/limited option is possible ? Because, for now, i feel it is too risky and won’t use it, despite it’s potential.

This is a good tool as a GUI onto the Whisper transcriber tool. If you aren’t confident with the Terminal version of Whisper then this presents a nice interface with a choice of models to use to transcribe audio to text. However, I was particularly keen to use it for identifying speakers from meeting audio and so I paid to upgrade for this feature. I’m afraid to say it’s awful. On one recording, it was two men and a woman and all at different volumes because they sat at varying distances from the microphone. It could not determine accurately any speaker and is thrown completely if someone says ‘yeah’ or ‘right’ which happens frequently in conversations. I’ve cancelled my subscription. So if you want a fast, user-friendly audio-to-text transcriber this is great. But if you want Speaker ID, it’s not there yet. I’d stick to the free version.

Obwohl behauptet wurde, dass ich meinen Lizenzschlüssel auf 3 Geräten gebrauchen kann, war es mir nicht möglich diesen Schlüssel bei meinem Laptop ein zugeben.

I tested it out and loved it so much that I went ahead and bought the full version permanently! 🎉 Just one tiny thing I wish for… Korean language support, please!! 🙏🇰🇷 Pretty please add it soon! 💖 Also, I downloaded it on my iPhone too — just wondering, does the purchase carry over between devices? 🤔 Would love some clarification! Thanks for the awesome app! 😄✨

+ It has history + free on-device transcribe (works great) + paid in cloud transcription (faster)

Very good speech to text especially in English. Well-integrated into the OS, The one downside of Whisper technology is that it probably won't improve much, as OpenAI has moved to proprietary models in ChatGPT 4. The app store version does not include direct dictation into apps. The version not on the app store does allow direct dication which is far more accurate thann the native macOS dictation functionality. If you do a lot of dictation, I'd recommend that version. The plus side of the app store version is that family sharing means that the whole family can enjoy good text to speech. In our case, our family is gradually moving to LInux though.

Ottima soluzione per affinare appunti di riunione documentazione ecc a partire a audio e trascrizioni o video

Great app! The best transcribe I have ever used (And you can buy the full app once instead of paying every month). But I think it would be great if it support multi-language transcription in one meeting (my team is multi-language and speaks English, Chinese and Japanese in the same meeting). Also, the auto-language detect seems to be weird, it keeps detecting strange language that no one used in the conversation like Russian. Also, I think this is the biggest fall of for me, is that this app doesn’t have the option to choose "Simplified Chinese (China)" and "Traditional Chinese (Taiwan). The only Chinese language in this app is simplified. This option is almost standard in every other transcription services. The accent of two languages are different so it impact the accuracy, and the output of the transcription is also different because the words are different. Another wish is the choice to choose auto-output format. The standard one is m4a, which is hard to just transfer and listen to on other devices that is not Apple. But despite all the small improvements it can make, this is still the best app I have used. Even beats out most of the online transcription services in terms of accuracy and ease of use.