Those who need regular transcription services often look for a likely impractical combination of high accuracy and low cost.
Scribie teases that possibility with a service that potentially offers a high degree of accuracy and one that is remarkably cost-effective.
Is Scribe just too good to be true?
Subscriptions and Prices
Scribie offers two transcription options; one delivered by AI and another involving humans, and each is priced accordingly.
The automated AI service costs 10 cents per minute and promises good sound quality between 80% and 95% accuracy. However, it reduces that accuracy to 80% for accented speakers, 60% for noisy audio or poor quality recordings.
The minimum bill is $1 even if your recording is less than 10 minutes long and transcribed via AI.
Recordings are processed within 30 minutes, excluding speaker tracking or time coding.
The alternative is the manual process which predictably takes much longer, up to 36 hours, and costs $0.80 per minute.
At eight times the cost, they predict 99% accuracy, but you’ll get an extra $0.50 each for strict verbatim speakers with accents, noisy audio, poor audio, and subtitles files.
And, very importantly, both services only handle English speakers.
All processing requires upfront payment, and this uses a credit system that requires you to preload with money. Credit can be added to the account from $10 to $15,000, and a common fund can be managed between multiple team members.
As an aside, when we first covered Scribie a few years ago, the prices were the same as they are now. So taking inflation into account, this service becomes cheaper.
The mechanism for uploading audio files and retrieving the transcribed results is a simple web page associated with the billing user account.
Files can either be loaded directly from a computer or come from many cloud storage providers and video sites, such as YouTube and Vimeo.
Once the files have been uploaded or transferred to Scribie, you will be given a quote for the cost of transcribing a file.
As part of our testing, we uploaded the beginning of a famous book spoken by an English actor and comedian. The clip ran for 2 minutes and 3 seconds, and the cost to transcribe it using AI was only $1, and to have it done manually was $1.64.
AI processing was much faster than the 30 minutes allotted by the system, and once the file has been processed, Scribie will provide an estimated accuracy.
You can download the text in a variety of formats, or you can use a web-based editor to listen back to the audio alongside the transcription.
This online editor is not an impressive part of Scribie. The web page it appears on has no horizontal scroll bar and leaves huge amounts of unused white space on either side of the work area.
For manual editing, Scribie uses a four-step process to take raw recordings and transcribe them using home-based freelancers to perform the transcription and review stages, followed by an in-house quality assurance team to proofread the work and ensure quality.
The mechanisms for handling the work are identical to the AI pathway. The only difference is that it takes longer to happen and costs more to have it done that way.
Assessing the accuracy of an AI system is one thing, but a solution involving multiple people is inherently flawed by definition. That said, Scribie has clearly created a tight looping mechanism in their workflow to get the best possible accuracy within the limitations of the audio provided.
And Scribie is confident enough in this process that they will reassess the transcript at no additional cost to the customer if the results are unsatisfactory upon delivery.
That may not be convenient for someone who works with tight schedules, but it shows a high degree of dedication to getting things right within the vagaries of human communication.
If super-accuracy is important to you, we’d use the manual service, because in our tests, the AI transcription has its limitations.
One of the clearest speakers we could find was Mr. Stephen Fry reading Harry Potter, and we let Scribie loose on a small snippet of that audio. For the curious, the book in question was not written by JK rolling’, as Scribe interpreted it. But this was a small improvement from the attempt a few years ago.
Then it made a complete mess of the first chapter of the book that the Dursleys called “Dales” and Dudley “double.” The last attempt turned it into Danes and deadly, though Dudley was right on one occasion.
In terms of accuracy, Scribie is getting better, but it’s not something we’d assume doesn’t take a lot of work to make it perfect.
Results may vary depending on the shot, regardless of brightness.
Make money with Scribie
One of the aspects of Scribie that differs from many other vendors is that people with English grammar and typing skills can sign up to transcribe with them and make some money.
How much money?
Payment rates range from $5 to $25 per audio hour. And, depending on where you live in the world, the lower pay would be considered below the minimum wage, and the top amount barely a king’s prize.
On the plus side, for those who stay with the service for more than a few hours, Scribie pays some bonuses.
What’s important to realize, though, is that you can’t realistically transcribe an hour of audio into one real hour. Transcription is performed in 6 minute blocks, and Scribie’s estimate is that the average speed is 45 minutes for that 6 minute audio portion.
That speed equates to 7.5 hours to complete an audio hour, or 66 cents per hour at the low rate, and $3.33 for those who earn the most. Some freelancers who work for Scribie can get 18 minutes per 6-minute block if they use the automated transcription and fix the issues afterwards.
However, as you perform these tasks, employment is not guaranteed, and pay is low for work that requires not insignificant skills. Unless you work in a criminally run company, jobs in English speaking countries, however simple, pay more than this.
The money earned is paid through PayPal and you have to compete with others for the average of 492 new files that are added to the system for transcription on average every day.
In short, unless you have unlimited free time and want to develop carpal tunnel problems quickly, this is not a career path we would recommend.
Scribie uses industry standard 256-bit SSL encryption on its website, which encrypts all communications between servers and browsers.
In addition, all transcribers are bound by a confidentiality clause about what they are transcribing, and because they only get 6 minute bits of the audio, they never have enough content to see the whole of what they are transcribing.
Files deleted by users are permanently deleted from the servers and the account can also be deleted if deleted.
The company will also enter into company-specific confidentiality agreements if necessary, but the service is not HIPAA compliant.
Scribie is a service that seems cheap at first, and it is if you only use the AI options. But for those who want more accurate transcripts, the manual option may be less beneficial if they have a poor recording of an Australian giving an open-air presentation at an airport.
The American nature of this solution shows that most of those who transcribe for Scribie are from North America and as such find the regional accents of English speakers from other countries, ironically including England, a challenge.
While Scribie claims to have expertise in British, Australian, Indian, African, various European and Asian accents, there is an additional cost and longer timeframes for doing so.
Due to the additional cost and time, we recommend that the service is probably only suitable for people based in the US. For example, UK residents should be able to find local transcription services that don’t charge extra for their local dialects.
For those in the US, this solution seems very affordable, especially if you have shots good enough to achieve acceptable accuracy with the AI solution.
Those who require the highest accuracy can take advantage of the manual service, which is also relatively inexpensive if you don’t incur any additional costs.
There are better services, but not many as cheap.