dsaffy

Yeah we totally agree. That's why we work on the end-to-end app, not on a single prompt. You pick what parameters become knobs in the frontend. So if you have a giant app with 10 parameters (say 5 prompts, 5 numbers), great, wrap those and they become knobs on our frontend. We override during the actual end-to-end testing execution kinda like a Statsig / Launchdarkly (only with typing).

Awesome work Cai and Andy. Do you have advice on structuring functions for best results? (eg, is it a good idea to give my agent access to things like a calculator "in case it needs it" or be careful about spoon-feeding it the tools I think it needs)

Just tried it out in my dummy application, but it throws an error when you provide a suffix... which is officially supported by their documentation. Hoping they support this soon.

https://beta.openai.com/docs/api-reference/completions/creat...

I understand the cartoony feedback. We're working on improving the quality of the avatars to be more adult / work focused. We do believe in avatars - they are one of the key points in our product - because we believe that being on video all day is soul crushing, but audio-only feels disconnected. The avatar provides privacy and personality at the same time. FWIW, you can turn on video if you want it too, we just default to the avatar.

Appreciate it!

The "where's the demo video" has been consistent feedback on this thread and we'll probably put something together.

Thank you :)

Thanks!

We have a lot of Japanese users (organically, over 25% of our users) so we internationalized.

ProductHunt leaderboard was a weekend project for me... I won't lie, there's not much connection, just some fun.

Thanks for the feedback.

Hey there - there is a search button that allows you to search messages. Is that not working for you? Reach out to us via the Pesto team chat in the app for support on this.

I am the author. FWIW I totally understand where you are coming from; I'm not happy with our explanation of "What is Pesto?". We will take our own pass on refining this, and I'd also appreciate if you tried it and gave me your thoughts on it. Right now "digital workplace" is the best we have.

We used Lottie

The core concept is "rooms" - like Discord voice channels, only way more flexible and built for remote workers. They can be traditional video conferencing, lofi spaces for hanging out, or 2d spaces with spatial audio.

The goal is to give teams a flexible workplace consisting of a series of rooms purpose-built for their workflows.

We are focused on rooms, which are audio-first spaces, not text conversations like Slack. Rooms are flexible containers that can be as simple as a traditional video conference, or more sophisticated, eg walk around on a map or listen to lofi music.

You can't look at screen share without permission - this is clearly something we screwed up in the landing animation. It's supposed to show Katherine sharing her own screen.

Pricing is coming soon and will free tier + paid tier (think like Slack / Zoom pricing).

There are parallels to Discord; the HN audience knows Discord well and so if we were only trying to appeal to this audience we would probably market it that way. A large chunk (possibly the majority) of our existing users don't know about Discord though and it doesn't make sense to purely market it in relation to that.

There are also some noticeable differences, particularly: - integrations are focused on work products, eg calendar, rather than what games you are playing etc - rooms (similar to Discord voice channel) can be shared on a calendar for guests - we prioritize different things in screenshare (quality over FPS, since it's for work not gaming)

We plan on introducing a paid tier soon. Think Slack / Zoom pricing, free tier + paid tier model. We haven't totally figured it out yet.

Hey Vincent - the GIF is supposed to show Katherine joining a room and sharing her own screen. You cannot force Katherine to share her screen.

You can manually set a status :)

Pragli is totally free until at least June 1 to be helpful during COVID-19 - we haven't determined our pricing model past then, but it will probably resemble Zoom or Slack (a free tier + a paid tier that's priced per user per month).

By default, the avatar grays out eventually if you don't interact with your computer, but otherwise is static. You can optionally instruct Pragli to use local facial recognition to move when you move and smile when you smile. Some users really like it and it looks more lively in demos, but FWIW the strong majority of our users are on keyboard/mouse. Here's a blog post I did a while ago about the facial recognition mode: https://pragli.com/blog/live-avatars-with-faceapi-js/

Awesome! Let me know how it goes.

Pragli is totally free until at least June 1. At some point, we will implement a free-forever tier + paid tier. We aren't sure what the pricing for the paid tier will be, but it will probably look like Slack / Zoom ($ per user per month).

Author here - we put up the logos for companies where members of the company use Pragli. We note in our terms of service that we reserve the right to do this, but we are happy to remove any logos of companies who request not to be used as such.

Re security: we do have e2e encryption; if you have a moment, take a look at our security brief under our FAQ https://pragli.com/blog/faq/ and let me know what you feel we should add / how we should display it.

Re Discord: I'm a gamer so I understand why you like it. Where I feel Discord is lacking in the workplace is that it only shows if you're playing a game, and it doesn't build any more sense of availability beyond that. It doesn't have the same auto-status / Spotify / manual status that we built into Pragli. It does for games (shows if I'm playing some game) but not for work.

Sure thing! A couple keys differences: 1) Avatars - as you mentioned 2) Walkie-talkie - most virtual offices have a call analogy, where you call someone to speak. In Pragli, you don't call - you just jump straight into a conversation, but the other side is muted (to protect their privacy). Here's my blog post on this: https://pragli.com/blog/building-a-walkie-talkie-for-remote-... 3) Some solutions have a very literal office (with a floor plan, etc) - we have a more virtual analogy 4) From what we hear, we have higher quality audio, video, screen share than others (I encourage you to evaluate for yourself)

Hi there - currently, we encourage you to use Pragli alongside a text tool (Slack, Teams, even just email). We will evaluate whether or not it makes sense to build one as we go.

Hi HN! After selling our first company, my co-founder Vivek and I started working remotely at the acquirer. We discovered we were much more productive than we'd expected, but we struggled with two main issues: communication and loneliness.

Communication friction came primarily from the endless back and forth in Slack / text communication. On the other hand, scheduling a Zoom meeting felt very heavyweight.

Loneliness came more slowly, but speaking to people less often definitely affected my happiness and productivity.

As a result, we built Pragli, a virtual office for remote teams.

In Pragli, you customize your appearance with an avatar and then can hop into audio rooms with one click. Video and screen share are optional.

When you're focusing or other unavailable, you can manually set a status, or automatically set a status via our Spotify and calendar integrations.

FWIW, TechCrunch describes Pragli as "BitMoji for the enterprise." https://techcrunch.com/2020/04/14/pragli-remote-office-appli...

You do have to sign up with Google, Microsoft, or Github OAuth, but there are no additional hurdles beyond that, you are dropped right into a team. You also have the option to skip downloading the desktop app and play with it in the web.

I'm here to answer any questions. Let me know what you think!

We actually discovered that the CPU increase was not caused by the detection, but rather by the CSS animation of the SVG moving around.

It turns out that the detection itself is very fast and lightweight, so there was no need to optimize (background workers would've been a good idea though).

We somewhat optimized the CSS animation, but it's still not where we'd like to be... if anyone on here knows how to improve CSS animation performance beyond dropping frame rate / shortening the animation, I'd love to know. Tried some obvious things (z index, will-change).

We are a bootstrapped 2 person company and the product is in beta. FWIW, the two of us devs are doing QA as well.