Record and get transcripts in real-time



You need to become a Contributor to join the discussion - Find out how.
Yoshio Goto@sushiogoto · Co-founder, Genie
This is really great! Could see myself using it in meetings where I want to be 100% present. I recently went to a doctor's office that had something similar — the doctor wasn't distracted taking notes which made it a much better experience. The only thing that would hold me back from using it is the privacy policy. I'd probably be recording sensitive information so wouldn't want my transcripts living in someone else's database.
Matthew WallerMaker@waller_matthew · Matthew Waller
@sushiogoto Thanks so much for the kind words! We hear you on privacy. We proudly use IBM's Watson for the raw transcription backend (there you have it, some of our secret sauce :) ). In interest of privacy, we opt-out of learning, and Blueshift doesn't hold on to any user data at all. This is from the Watson developer docs on how they handle the information they receive: "By default, IBM logs each request to the service and its results; IBM uses the data only to improve the service's base speech models for future users. You can prevent IBM from storing your data by setting the X-Watson-Learning-Opt-Out request header to true for all requests. When you use the header to opt out of request logging, IBM stores none of your data; your data exists within Watson only while it is in transit (while the service is processing your request). In either case, your data is always encrypted both in motion and at rest." Source -
Yoshio Goto@sushiogoto · Co-founder, Genie
@waller_matthew thanks for the thorough response! Will be testing this out in an upcoming meeting :)
Yury Molodtsov@y_molodtsov
What is the difference comparing to Cassette and other similar apps?
Matthew WallerMaker@waller_matthew · Matthew Waller
@y_molodtsov Hi Yury! We have a few differentiators. 1) The biggest one is that, with Steno, you can edit your transcript within the app. You can tap on a word and jump to the audio to easily see if it's correct. And you can tap on the waveform to jump to the right text. And if the text is incorrect or needs formatting, you can change it then and there, and still maintain the correct timestamps for all of the other words in the transcript :) Having in-app editing as well as the live transcription is a solid combination. 2) Our pricing is not tied to a subscription. We personally don't like dealing with subscriptions in app, so we let the users have more control over the number of minutes they want to buy. 3) And we have extra conveniences: in addition to standard importing, we have auto-uploading to third-party services, what we call syncing, and you can slowing down sound while you're listening without changing the pitch, and we have a clean and simple means of bookmarking. All of these things tied to one package is what sets us apart.
Nate Benson@natebenson · Digital Media Manager at 43North
I'm a Digital Media Manager at 43North in Buffalo and my background prior is in news. This could be a great everyday tool for journalists and video producers. Logging video is a royal pain and very time-consuming. The bookmarking and edit feature is fantastic. I think you could really market this to news organizations and video producers. If you went that route, one thing I'd recommend is being able to enter a manual timecode prior to the edit. That way when the editor gets in the cutting room they can quickly pull up timecode in their NLE and correspond that with bookmarks (which would also lock a specific timecode of a sound bite they liked while recording). If you marketed the app to that industry you could really get some traction. I'd even be willing to pay for such an app, between $2.99-4.99 easily if it was a near perfect transcription. I will admit that some of the video I played through speakers and transcribed missed a few words, but the overall gist of the sound was intact. But adding some features that would cater to content producers would make this a must-have app in newsrooms, digital agencies, etc.
Joel Farris@jifarris · Founder, Return Technologies
I just tried this out by recording audio from a podcast and its accuracy was impressive! I like how the sentences are broken up. And being able to jump straight to different parts of the recording via the text itself makes correcting errors super quick. I’m thinking podcast, video, meeting transcriptions… I’d be interested to hear how others might use it.
Mike Irene@michaelirene · VP, Software Eng, PRA Health Sciences
Accuracy seems pretty good. I can see potential for conference call transcriptions. Miss a meeting? Have someone Steno it ! But it needs one important enhancement -- the ability to identify and annotate who is speaking. I can envision that you setup your own voice (or colleagues) similar to Siri so it can recognize you vs others or just mark as Speaker1, Speaker2, etc. You manually edit on playback and it cascades through the transcript.
Mark B.@sbmarkb · Product Manager, CJ Affiliate
@michaelirene This is exactly my thought on the next "big" feature... even a rudimentary voice detection implemented to identify voice1, voice2, voice3 with the ability to "name" the voices afterwards would massively increase the value.