Anyone heard of Speak.ai? Text-to-speech, sentiment analysis, and now Evernote integration

I recently learned about this new startup that’s doing text-to-speech, similar to Otter.ai, but they also handle video, and have a built-in note/document creation system as well.

It seems like they want to sort of be a central repository for all your notes, whether audio, video, or text. The kicker is they do sentiment analysis and other AI-driven parsing of the text in the system, whether from text-to-speech output, or direct writing input. It’s a really interesting idea, and while the initial implementation is a bit rough so far, there are some cool features, like the ability to create an embeddable recorder to collect audio or video from people, feeding it into the system, then doing text-to-speech and sentiment analysis, etc.

Today I got an email announcing new Evernote integration, and while I don’t use Evernote myself, as a quantified self enthusiast one thing about the integration caught my eye, highlighted in the screenshot below.


Is this possibility exciting to anyone else? Have any of you tried Speak.ai for yourself? What do you think of it, or at least what it’s trying to do? Do you see a place for this in your toolkit in the future, if it improves enough?

2 Likes

This is certainly an interesting find !
I could easily see myself using it as a part of my workflow , if and when :

  1. They bring in more integrations - to the level of readwise
  2. Who am I kidding ? I am definitely trying this one out !

Will get back to you about my thoughts on Speak , once I have had a chance to try it out !

First impressions :
- It looks like they are taking it a notch above Otter.ai . They can transcribe audio as well video , analyse them as well as text and give you detailed reports as well ! This is still free :wink: I am in :ghost:

1 Like

Yeah, it’s definitely interesting at this early stage! I do think I’ve run into limits on what/how much you can transcribe from audio/video. And they have a pricing page which says there should be:
https://speakai.co/pricing/ (“free minutes”)

But things seem a little buggy as far as whether those limits work, or how they’re presented to the user.

Also looking at that page, you can see perhaps some of the features they’re planning. They seem to be taking a lot of publicly available deep learning-type tech (either as API, open source, or published research) and basically just throwing all the content they can at various engines to see what stuff they can get out. For example celebrity recognition? How large a market does that seem like it might serve in this context? But hey, if the tech is there and the source data is there, why not try it? :man_shrugging:

Regardless I think there’s enough there to play with and maybe get a sense of whether it might be valuable to you. The text analysis - rather than just transcription is certainly an interesting idea anyway (and similar to some things Codex developer @argimenes has mentioned on Twitter, e.g. sentiment analysis of text in the system).

1 Like

Yes I believe it’s already in place . But unsure how effective it is , beta testing would say I guess

1 Like

I forgot to ask, do you use Otter.ai already @Karthikk? If so, for what?

I do use it occasionally, but used to use it more. I have a daily logging habit that I’ve been keeping at for almost a decade now, and I find it quite valuable. But it is of course quite hard to maintain on mobile or any time I’m away from a computer! I use Otter to record audio versions of my logs (abbreviated). It works OK, the transcripts give me a head-start on meaning, but ultimately I need to transcribe myself for best results.

Sadly no . I find it very comfortable to dictate my notes through drafts app into my Apple Watch .
It does a really good live transcription , as well . Unfortunately there’s a time limit for a single dictation and that’s my annoyance for now . If and when Otter.ai bring out a watch app , I will definitely try to incorporate it into my workflow .

1 Like

Ah, same for Android. The “voice typing” works pretty well, but it times out. Hence Otter.ai. :smiley: But ultimately I want it all in text, it’s most searchable, linkable, accessible that way. The quest to get things out of my head and into text most efficiently is ongoing…

1 Like

Maybe we dont have to transcribe it all ? Annotating any media form is what codex is promising to be able to do , and then link it with the entire knowledge base .

Fingers crossed :crossed_fingers:

1 Like

Right, but in general it takes more time to listen through the audio later to even make connections, much less review or reference it. So transcription is still ideal, even with audio annotation. Pretty excited about the potential in “annotate anything” though, to be sure!

1 Like

Oo, I just found their public roadmap:

And for anyone interested in a deeper dive into their progress and plans as a company, they just put out a fairly detailed quarterly report:

(you can enter any email to get access, even an invalid one, shhh! :grin:)

1 Like

Ok, what got me excited the most is the fact that I am able to access the Trello board directly from the forum .

Haha , highly valuable information :joy:

2 Likes

Yeah, apparently some of the Discourse embeds are “rich”, i.e. interactive! Pretty cool. I wasn’t able to find a list of which sites are supported that way though…

1 Like

Hey everyone :heart_eyes:

I’m so excited to find this “little corner of the Internet”.

Thank you so much @Oshyan for connecting and for sharing the work we are doing at https://speakai.co.

We are truly passionate about it. Sometimes we find it a bit difficult to communicate what we are trying to do and then we get introduced to a place like this that feels just like home.

I really appreciate the dialogue in here. There are a ton of things I’d love to contribute but for now, I just wanted to say hi.

Thank you again. :hugs::heart:

Please feel more than encouraged to connect with me!

All the best,

Tyler Bryden

https://www.linkedin.com/in/tylerbryden/

2 Likes

That’s exactly why we wanted to foster a community where makers and users could connect with each other , understand and iterate :slight_smile: Obviously once the tool becomes matured enough they are bound to have their own community , but for the one’s that are still in their nascent stages , trying to find their market niche and understand user emotions ? Those are the ones that would benefit the most from our community :slight_smile:
Thrilled to have you here !

Also , @Tyler_Bryden this forum was launched pretty recently , we have a larger discord audience here https://discord.gg/Axd4BPCF. Feel free to jump in anytime and introduce yourself as the founder of Speak.ai, we have quite a few other founders there as well .

1 Like

Hey @Tyler_Bryden, sorry about the very late reply, wow. I haven’t really had much chance to use Speak Ai since I posted this, unfortunately. But I do keep an eye on your newsletters and the pace of new features and enthusiasm you bring to the product are evident and really appreciated. I hope you’re finding some traction and success!

You’re definitely trying to do some very interesting and unique things vs. incumbents like Otter.ai. It does seem as if there is a bit of an exploratory or at times less focused nature to your approach, but I imagine this is because it’s pretty early in the product and there are lots of ideas to try. I’m hopeful you find market fit through some of your experiments, and I’m curious to see what that will end up looking like.

The sentiment analysis, etc. are of interest to me as a quantified self enthusiast, but I’m never clear just how big a market that is… I think it could be a bigger one though if the tools to do such self analysis were easier to use, faster, and more transparent. The fact that most people that are into it at this point are a bit hardcore has a lot to do with the available tools and integrations, I think. But there are newer apps trying to streamline things… which makes me wonder if you’re planning integrations with some of them, actually.

Anyway, bit of a ramble, but I wanted to keep this discussion alive and show some appreciation for your work! :slight_smile: