Sean-Der

I see a couple ways to combat that.

* Your TURN server should provide APIs that allow you to verify that allocations/permissions are only created for your users.

* Use an auth mechanism that has an expiry time. Like [0]

[0] https://github.com/pion/turn/tree/master/examples/lt-cred-ge...

For me something only qualifies as a 'huge ugly hack' if a better way exists. I think TURN does the best it can in the constraints that exist!

* TURN handles protocol bridging (TCP <-> UDP)

* TURN is useful for privacy preservation.

* TURN handles NAT Traversal (unfortunate when it comes to this)

* TURN is used for security in some cases. It sits at the edge and clients create allocations

Great write up! It isn’t free, but the price/what it offers is really great with Hetzner. I switched from digital ocean.

Did you evaluate any other TURN servers? I’m curious about your thoughts of the Elixir and Go ones. Maybe even more exist, I haven’t looked recently

I’m so excited for this. If we get better protocols I anticipate so many more projects will get built. Check out this [0] article to get more on what WHIP solves.

https://github.com/Glimesh/broadcast-box is also a decent jumping off point

[0] https://webrtchacks.com/webrtc-cracks-the-whip-on-obs/

> also we noticed if you pass an empty tools array the api silently crashes

I will go look into that!

> Also would be super cool to just allow direct sip connections

Soon hopefully! We have all our own SIP infrastructure (from 1800-ChatGPT). Just need to expose it as an API.

People much smarter then me are working on the hallucinations. It's a pretty interesting world writing evals and improving the models.

If anyone has questions about WebRTC + Realtime API I would love to answer them. I am incredibly/excited about the IoT/Embedded side of it especially :)

You can see a video demo of it here[0]. The code on microcontrollers is IMO delightfully easy to get going on [1]

[0] https://youtu.be/14leJ1fg4Pw?t=824

[1] https://github.com/espressif/esp-webrtc-solution

You can! https://github.com/pion/offline-browser-communication

It uses mDNS for discovery. You can only do one browser though.

I don’t see a path forward on browser/browser that is obvious. It would make it so easy to fingerprint if you could set your mDNS hostname in JS

Email me sean@pion.ly

If you can work/create an interesting Open Source project I can send it around and get you some attention. That is a good way to meet people (and find leads).

If you are interested I could also collect some money to pay you to work on Pion! I have plenty of small tasks. We could apply to NLNet for those tasks. We could also start a OpenCollective and cut you up a list of tasks.

I’m very curious about what a LLM could deduce if you sent in lots of sensor data.

I love my Airthings. It don’t know if it’s actionable, but it would be cool to see what conclusions would come up from sending co2 and radon readings in. Could make understanding your home a lot easirr

I will work on making it better! This was announced Tuesday [0] I still need to give it lots of love.

Even though the README isn’t completely done, give it a chance I bet you can have fun with it :)

[0] https://youtu.be/14leJ1fg4Pw?t=625&si=aqHm1UAdDEz91TnD

Thank you! If you run into problems shoot me a message. I really want to make this easy enough for everyone to build with it.

I have talked with incredibly creative developers that are hampered by domain knowledge requirements. I hope to see an explosion of cool projects if we get this right :)

Sorry I forgot to put use cases in! Here are the ones I am excited about.

* Making a toy. I have had a lot of fun putting a silly/sarcastic voice in toys. My 4 year old thinks it is VERY funny.

* Smart Speaker/Assistant. I want to put one in each room. If I am in the kitchen it has a prompt to assist with recipes.

I have A LOT more in the future I want to do. The microcontrollers I was using can't do video yet BUT ESP32 does have newer ones that can. When I pull that I can do smart cameras, then it gets really fun :)

Lots of custom logic

* Limit user to 15m a month * Greetings unique to user state

It also needs access to Model/DB etc… which is all not exposable to internet

Thank you! I didn't feel any disrespect :)

I was always curious how things worked when I saw announcements on HN. So happy to share to satiate the next generations curiosity :)

No teleprompters!

We had a video monitor in front of us showing the live feed, that kept distracting me personally.

I worked on this.

My background isn't AI so I can't contribute to that. My background is WebRTC/telephony so I could build this. Even if I was involved in 'AI stuff' I would have zero impact, but I can build this!

Yep! Twilio is providing the SIP Trunking. We are running our own SIP servers though.

You can build it now! https://github.com/openai/openai-realtime-embedded-sdk

Here is a demo of it https://youtu.be/14leJ1fg4Pw?t=805

Not outsourced, I worked on it myself.

This is the old owner of the number. Some carriers are still routing it wrong.

Peer calls can run in P2P mode or as a SFU.

Different topologies are better for different use cases. 1:1 vs group calls in this case!

I think people are excited because it’s easy to run yourself and well documented.

I find the code base easy to understand since it’s built by one developer. It’s a consistent read and not sprawling/pointless duplication

My dream would be if WebRTC Agents would experiment with Port Control Protocol.

I don't know if it is the answer, but a world without dependence on STUN servers sounds pretty amazing. I can see why it would feel like kludge, but how could would it be to remove that dependency.

My other wishlist item is to allow mDNS for signaling. Something like https://github.com/pion/offline-browser-communication. It is so silly that I need to exchange a Offer/Answer using a remote server to start a session with a IP Camera in my LAN.

I don't think Chrome or FireFox support passive ICE-TCP Candidates.

The assumption is that two clients are in the same LAN they would use UDP anyway. If are in two differents NATs doing traversal with TCP isn't support. I have read that it is possible, not something I have tried myself though.

You should build that!

Check out the initial commit of https://github.com/awslabs/amazon-kinesis-video-streams-webr... it was Pure C and simple.

Early versions of https://github.com/pion/webrtc used lots of cgo its an enjoyable read I find. The glue is higher level (interesting part) but the parts you don’t need the details on (DTLS, SRTP….) are abstracted away

I love that WebRTC re-uses existing IETF standards. Makes it easy to plug existing pieces together.

I tried to capture that with webrtcforthecurious.com breaking it down by subsystem.

I should do a companion piece ‘Build yourself a WebRTC’ that shows how to build WebRTC yourself in Python or other popular language.

ONVIF is getting WebRTC [0]! So you will be able to use it in your LAN (or over ) without using proprietary middleware :)

[0] https://www.onvif.org/specs/stream/ONVIF-WebRTC-Spec.pdf

I’m working on this!

A WebRTC implementation for my microcontrollers exist https://github.com/sepfy/libpeer it requires more work to use then most want (but it is the perfect building block for higher pieces)

I am trying to make it easier with a SDK for LiveKit https://github.com/sean-der/embedded-sdk

My assumption is the amount of people encoding is much smaller then decoding.

overlap exist in encode/decode. When working on a Opus codec I did have a fair amount of code distinct to one path still.

It’s my passion in life to see WebRTC utilized to its full potential! When I learned about P2P/NAT Traversal I could never go back :)

Why do you think it’s had trouble? Poor software, educational materials etc… always looking for new projects