This new Hermes Desktop browser isn't just an extra window with a browser bolted on, the agent can operate, see, and analyze what it sees along with you I made a short video demonstrating this, but here are just a few things you can do with it: You can... scroll X, then ask for a summary of your feed or what the hot topics are You can watch a tutorial video from your favorite YouTuber **cough cough* and then ask your agent to pull transcripts, explain certain concepts from the video, or even implement them directly! You can ask for research on some open source memory solutions for agents, open those links right in the browser, then review them togethershow more

Sweet! iOS and Android simulator next? @Teknium is that a possibility?

how can i use it？it's a plugin？

I don't believe it is a plugin, just an in-app feature. Just ask your hermes to do it!

@Teknium Can this be done remotely? Or only by seating at the computer using desktop app. I want my agent to see and read but remotely so I don’t have to be at my desktop

@Teknium So it does appear that it needs to be with a local gateway

guys I know it seems a dumb question but : HOW TO OPEN IT ?!?! I've even asked to hermes but even him seems confused :\

You need to be updated to the latest version, and be running on a local gateway for it to work apparently. I simply asked hermes to open the in-app browser like you saw in the video

seeing the actual screen beats scraping the DOM every time. half of X context is visual, so an agent that watches along is way more useful.

Sadly can't use with my remote gateway

Yes it appears to only work on local gateway for now

@tonbinstudio this won't work for remote gateway right?

No it seems like it needs local gateway

@Teknium So cool hey.

Is it possible to have Hermes running on VPS whilst having desktop on your main computer connecting to it? So all work is kept on the VPs and the desktop is simply a window to it? @witcheer

Looks like I won’t need to keep a separate X window open anymore.

Hi Tonbi, question. I was following your instructions and I got the browser to work. One thing I noticed was that my desktop was a lot slower than yours. You were getting it to open in several milliseconds. I was on the order of 5 to 10 seconds. I suppose it could be machine speed because I'm running it on a Mac mini base, but I'm wondering if there might be other factors related to load or context. I want to speed up my desktop a little bit. It feels a little slow to me and sluggish at times. While I'm waiting to buy my super machine, I wonder if there are any tips that I can do locally to help. On a separate note, I couldn't get the new voice conversation mode working properly. It was very, very laggy, nothing like my codex, which is a quick back-and-forth conversation. Any tips about that? did you do a video about that also?

Great! In terms of your desktop speed, it could just be based on the memory usage/speed of your machine. Maybe the model too, I was using a faster model, maybe a slower one just took longer Voice conversation is fairly new and I haven't really tested it all that much, I'm more of a text guy, but I'll take a look

Thanks for the quick response. I enjoy the videos. Yeah, I definitely think making that model at high versus fast should be investigated because that might be the trick. Thanks for the idea. Regarding voice, I've recently started using voice more and more, and I really like it. I love it in codex with the conversation.Haven't been able to replicate the codex experience yet, despite using fast models for that. I found one or two bugs in the voice files also that I reported,so I'm holding off a conversational voice in Hermes until I see those are resolved. Wispr Flow is working fine for me.

Following up on this I had switched to Fast model and I'm still finding it very laggy, nowhere near my experience on Codex. In fact when I use Codex Fast, it's great. It's like talking to a friend. When I use Fast modes in 5.5 or 5.6 GPT via OAuth in Hermes, it still feels very laggy and it still prints out the call response and then starts talking

@Teknium Anybody remember those 2 guys who released the brainrot attachment for cursor so you could watch YouTube or gamble on stake while in cursor all one window. This is like the productive version of that, @NousResearch are making incredible strides in so many directions

that command doesnt work for me. i dont see this anywhere in the desktop ui, and my agent doesnt seem to be aware of its existence. how do i enable this? im running hermes via gateway if that makes a differnece

You need to be using the local gateway I believe

Hermes is shipping crazy features 🚀

That's cool...but how do you have the tabbed sessions there at the top of Hermes Desktop?

They shipped tabbed sessions before, it is just a normal pane

nice

Amazing

life-changing this browser thing is it won't ban me, right? using to access linkedin posts / x timeline etc

I haven't been banned yet!

tried yesterday - turned my 1 hour work in 10 min.

Right?! Seeing the actual screen is legit everything. DOM scraping guesses. Watching along is how you catch the stuff a text summary just... skips.

Much needed shiii

Is this Mac only? I don't have it and I'm updated on PC.

Hermes-verse day 62…I have Hermes open watching me watch this, watch yourself, watch Hermes scroll your X feed while I’m sitting here in my Hermes X feed….sendhelp 🫠😂

This is all too much lol

An agent that actually sees and reacts to the page is a real step up

is there a way to get an address bar, forwards/back buttons, open the browser without asking the agent to do it for you?

Will try this out.