Send e-mails verbally with Jott

I’ve been trying out this new service called Jott that converts your voice into e-mails and text messages. Once you sign up for a free account — a pain-free process — you can call a toll-free number to create a message. You’re asked who you want to Jott to and you can say “me” or “myself,” or the name of any of your contacts (which you can import via Gmail, Yahoo, CSV, etc). After you say the message, the system says, “Got it!” and you hang up. A few minutes later the message is transcribed and you or your recipient receive it via e-mail. You can also set reminders to receive a text message alert at a later date and time.

Since I bought my BlackBerry in November, I’ve been using Simulscribe’s visual voicemail and now almost never actually listen to my messages, but read them instead. It’s faster to read than to listen, and I can do it while I’m in a meeting or on the subway. Their transcription services are at about 80% accuracy, which is fine for my purposes. Most of the time it’s just proper names that they get wrong. Plus it only costs me $10 a month and I enjoy the futuristic-ness of the service.

When I heard about Jott I wanted to see if their accuracy was nearly as good and wanted to further explore what the benefits could be. I did a test last night and this morning and the results are below. Each screenshot is the transcription sent to my e-mail inbox.

Message 1:


I’ve highlighted the message to show you the inaccuracies. I had actually said, “…and it’ll be transcribed when I’m done,” but it heard “they’ll” and “transcribe.” I have to be fair and disclose that I’m a New Yorker and probably slur my words. Still, more than the gist of the message is there. If there were any confusion, I can listen to the audio message by clicking the link at the bottom of the e-mail.

Message 2:


I had said “Greg,” but it didn’t hear me so it put an ellipsis in its place. Surprised that such a simple name wasn’t recognized, and that the transcription confidence in the upper right corner still said “High” even though it knew it had missed a word. Still, I know that there’s only one package I’m supposed to take to the post office today. No real harm done.

Message 3:


What I had actually said was, “Clean and Clear,” a brand name which it obviously didn’t recognize. I imagine that the attempted translations could be a whole lot funnier than this mistake.

Message 4:

Here’s where I got creative. I thought I would try to stump Jott by singing one of my favorite songs into the phone. Absolutely shockingly, it got all of the lyrics right, even the capitalization of proper names. Unfortunately it cut me off before I could finish the verse. So maybe messages are capped at 30 seconds? Still, an incredibly impressive performance, especially considering my terrible voice. Could they have a database of songs pre-transcribed?


My Jotts

All of your Jotts are accessible online in an inbox interface. You can edit the transcription, rate its accuracy, organize them into folders, or add a priority. They also have text-based Jott entry so that you can manage your to-dos in one area without having to always call it in.


The interface is pretty simple, but the iframe scrollbars are atrocious. I don’t know if it’s a browser issue (I’m using Firefox on a Mac), but that has to be fixed.

Jott Links

Jott Links allow you to Jott to third party services such as Twitter, Google Calendar, Amazon, Blogger, Remember the Milk and more. Once you set up the link, the recipient of your Jott is the name of the service.

I’ve only tested it out with Twitter. Called Jott, at the prompt said “Twitter,” and then spoke my tweet.


I had said “tweet” but it heard “Twit” instead. Maybe they ought to add tweet to the Jott dictionary. Nevertheless, still a good showing and a great way to tweet on the go. Your voice message is even linked from the tweet so followers can hear it. The biggest problem will be knowing when you’ve exceeded 160 characters.

Overall, I think it’s an interesting concept and the implementation is better than I expected. Still, for someone with a smartphone it doesn’t have much of an added benefit; even on the go I’m able to use my keyboard to send an e-mail, write a to-do item or tweet. Jott makes a point of saying that it’s a safe, hands-free way to jot things down, but given that I don’t have a car it isn’t a necessity for me. I’ll use it for its novelty factor, so long as it remains free.

I think the biggest benefit of trying this out was discovering all the “Web 2.0” productivity services that they link to. I’ll have to go explore those now…

Related Posts:


  1. says

    Once you’ve been accumulating a lot of jotts, the inbox starts getting cluttered, and I don’t see an easy way of multiple selection and re-filing and archiving jotts to clean up the inbox. Their drag and drop method is nice idea, but limits you to one item per click with the associated HTTP round trip delay.

    They could use some keyboard shortcuts.

    I think has one of the snappier web UIs that deals with lists.

  2. says

    Case, you’re right. The fact that there’s no multi-select is just plain wrong. It’s an inbox. You should be able to take batch actions.

    Jott actually just redesigned their interface. Read more about what’s new here:

    And yes, Remember the Milk rules. I use it for all of my to-do lists. I had been using 37 Signals’ Ta-da Lists, but was thoroughly annoyed by the absense of a due date for each task. Then my manager Josh recommended RTM. I’m much happier with it, and their mobile site isn’t bad either.

Leave a Reply

Your email address will not be published. Required fields are marked *