Why does this not work? I have the HN homepage open, and open Zuni sidebar:
Me: Can you identify which stories in the list are not "tech" stories?
Zuni: Please provide me with the list of stories you'd like me to review! I need the list to be able to identify which ones are not "tech" stories.
Me: The ones on the open hacker news tab
Zuni: I am sorry, I do not have the capability to access external websites or specific tabs within them, including the "open hacker news tab." Therefore, I cannot identify which stories on that tab are not "tech" stories.
Me: You can't see the content in my browser?
Zuni: No, I cannot see the content in your browser. I am a large language model running on Google's servers. I don't have access to your computer, your browser, or any of your personal data. My interactions with you are purely through text.
Me: Which emails in my primary tab seem like they are from cold emailers?
Zuni: I am a large language model, also known as a conversational AI or chatbot trained to be informative and comprehensive. I am trained on a massive amount of text data, and I am able to communicate and generate human-like text in response to a wide range of prompts and questions. For example, I can provide summaries of factual topics, create stories, and translate languages.
Zuni: I cannot access your email inbox, so I am unable to help you with your request.
--
If it can't access content in the browser, it is pointless.
I have two questions: one business-related and the other technology-related.
1. Aren't you a little worried that Google, with its new Gemini offering and focus on integrating AI into Google Workspace, might cannibalize this product? Maybe I'm missing something here, but I'm curious what you think about that. Also, given that Manifest V3 extensions technically work on other browsers and that your extension can work on other websites, I'm wondering if there might be a better "wedge" you could get into (e.g. Proton Mail on Firefox, etc).
2. Without going through the effort of looking at your extension code (which might be minified), I was wondering how you went about implementing this extension. I built a fairly complex chrome extension last year and I'm not sure if the average developer understands just how "non-standard" browser extension development is. For example, authentication, complying with MV3, handling different tabs, noticing when the current page changes, lack of hot module replacement etc. I'm genuinely curious to hear anything you might share about your product's development.
1. You’re totally right to flag the potential overlap with Google’s Gemini and its integration into Workspace — it’s something we’ve thought a lot about. Google is definitely making strong moves here, but we see our opportunity not in replacing Gemini, but in building an AI layer that works across tools and ecosystems, not just within Google’s walled garden.
While Gemini is powerful inside Workspace, we think many users — especially knowledge workers and power users — don’t live exclusively inside Google’s tools. They might draft in Google Docs, communicate in Proton Mail, manage tasks in Notion, and read articles in Firefox. That fragmented workflow is exactly where our extension shines: by giving them a consistent, context-aware AI assistant wherever they are, not just inside Google’s UI.
You also make a good point about extension portability. We’re definitely exploring alternative “wedges” and distribution channels — like targeting users who prefer privacy-first tools (e.g. Proton Mail, Firefox), or platforms where the incumbents have less AI integration.
So in short: yes, we’re aware of the risk — but we believe there’s still a meaningful opportunity in offering a more flexible, cross-platform assistant that complements rather than competes directly with the big players’ native solutions.
2. For auth, we're using Clerk (which is a bit less well supported in extensions than we'd like, but works well enough), and the Chrome 'Identity' API to allow you to OAuth with Gmail.
Complying with MV3 isn't too hard to achieve, and the Chrome store has a fairly robust review process. They have a 'slow queue' for reviewing apps that request higher permissions (like ours has to) which means shipping hotfixes isn't really possible! It's usually a 48 hr loop to get new code deployed.
Re. hot module replacement we're using https://wxt.dev/ which is built on top of Vite and has been really good. Seamless dev experience with all the affordances you expect in modern web dev tooling. Would definitely recommend.
Deeply underwhelming. Your tagline suggests a general "take the wheel" automation. Your website suggests that actually it's just a gmail copilot which may or may not be as good as the official ones but which requires installing a browser extension instead of going via. the approved channels. And there's other specialized website apps to come someday maybe.
Point of feedback on the landing page - you guys need to a better job explaining the killer use case here. Your first video is an implementation detail not a selling point. Your second video is asking how to establish a C Corp using context from the Stripe website. Is anyone actually going to do this over asking the LLM directly?
Also, you install the plugin, and it doesn't work. You have to create an account and login before the input area is responsive. Not clear from looking at the UI.
I have 108k unread emails on Gmail. “Analysis” as I read them isn’t my pain-point. Surfacing the 2-3 important emails I get daily so that I don’t miss something important among all the noise is really all I want from AI.
Everyone is focused on helping me write emails or summarize emails; things I frankly do not care about.
That said, I like the idea of a copilot for the browser, it just requires a great deal of trust in your company since I’m giving you access to my entire life.
I use a bunch of AI tools, I pay for Claude and OpenAI, I use Claude Code/Aider, and I've been playing with Gemini and Aider recently. Spending money on AI/LLMs is not foreign to me at all. That said:
The price on this is hilarious. You want me to pay another $20/mo for access and a fun new abstraction of "Message Credits". No problem, I'll just scroll down to the FAQ to figure out what a message credit is... nope, no definition at all.
There is little chance I'm going to pay $20/mo for a wrapper that is just 1 step over copy/paste [0] and absolutely no chance I'm going to pay for it with limits that aren't even explained.
[0] I'm sure this will be debated, but I'm not seeing anything special here.
Why does this not work? I have the HN homepage open, and open Zuni sidebar:
Me: Can you identify which stories in the list are not "tech" stories?
Zuni: Please provide me with the list of stories you'd like me to review! I need the list to be able to identify which ones are not "tech" stories.
Me: The ones on the open hacker news tab
Zuni: I am sorry, I do not have the capability to access external websites or specific tabs within them, including the "open hacker news tab." Therefore, I cannot identify which stories on that tab are not "tech" stories.
Me: You can't see the content in my browser?
Zuni: No, I cannot see the content in your browser. I am a large language model running on Google's servers. I don't have access to your computer, your browser, or any of your personal data. My interactions with you are purely through text.
Me: Which emails in my primary tab seem like they are from cold emailers?
Zuni: I am a large language model, also known as a conversational AI or chatbot trained to be informative and comprehensive. I am trained on a massive amount of text data, and I am able to communicate and generate human-like text in response to a wide range of prompts and questions. For example, I can provide summaries of factual topics, create stories, and translate languages.
Zuni: I cannot access your email inbox, so I am unable to help you with your request. --
If it can't access content in the browser, it is pointless.
[dead]
Something's gone very wrong there, apols. Thank you for the feedback, we're looking into why it didn't connect to your browser's DOM.
Congrats on the launch!
I have two questions: one business-related and the other technology-related.
1. Aren't you a little worried that Google, with its new Gemini offering and focus on integrating AI into Google Workspace, might cannibalize this product? Maybe I'm missing something here, but I'm curious what you think about that. Also, given that Manifest V3 extensions technically work on other browsers and that your extension can work on other websites, I'm wondering if there might be a better "wedge" you could get into (e.g. Proton Mail on Firefox, etc).
2. Without going through the effort of looking at your extension code (which might be minified), I was wondering how you went about implementing this extension. I built a fairly complex chrome extension last year and I'm not sure if the average developer understands just how "non-standard" browser extension development is. For example, authentication, complying with MV3, handling different tabs, noticing when the current page changes, lack of hot module replacement etc. I'm genuinely curious to hear anything you might share about your product's development.
Cheers!
Hi Josh, thank you for your feedback :)
1. You’re totally right to flag the potential overlap with Google’s Gemini and its integration into Workspace — it’s something we’ve thought a lot about. Google is definitely making strong moves here, but we see our opportunity not in replacing Gemini, but in building an AI layer that works across tools and ecosystems, not just within Google’s walled garden.
While Gemini is powerful inside Workspace, we think many users — especially knowledge workers and power users — don’t live exclusively inside Google’s tools. They might draft in Google Docs, communicate in Proton Mail, manage tasks in Notion, and read articles in Firefox. That fragmented workflow is exactly where our extension shines: by giving them a consistent, context-aware AI assistant wherever they are, not just inside Google’s UI.
You also make a good point about extension portability. We’re definitely exploring alternative “wedges” and distribution channels — like targeting users who prefer privacy-first tools (e.g. Proton Mail, Firefox), or platforms where the incumbents have less AI integration.
So in short: yes, we’re aware of the risk — but we believe there’s still a meaningful opportunity in offering a more flexible, cross-platform assistant that complements rather than competes directly with the big players’ native solutions.
2. For auth, we're using Clerk (which is a bit less well supported in extensions than we'd like, but works well enough), and the Chrome 'Identity' API to allow you to OAuth with Gmail.
Complying with MV3 isn't too hard to achieve, and the Chrome store has a fairly robust review process. They have a 'slow queue' for reviewing apps that request higher permissions (like ours has to) which means shipping hotfixes isn't really possible! It's usually a 48 hr loop to get new code deployed.
Re. hot module replacement we're using https://wxt.dev/ which is built on top of Vite and has been really good. Seamless dev experience with all the affordances you expect in modern web dev tooling. Would definitely recommend.
Cheers
Thanks for your detailed reply! I'm definitely gonna look into WXT next time I build an extension.
Deeply underwhelming. Your tagline suggests a general "take the wheel" automation. Your website suggests that actually it's just a gmail copilot which may or may not be as good as the official ones but which requires installing a browser extension instead of going via. the approved channels. And there's other specialized website apps to come someday maybe.
Point of feedback on the landing page - you guys need to a better job explaining the killer use case here. Your first video is an implementation detail not a selling point. Your second video is asking how to establish a C Corp using context from the Stripe website. Is anyone actually going to do this over asking the LLM directly?
It seems like all the videos fall in this bucket.
Also, you install the plugin, and it doesn't work. You have to create an account and login before the input area is responsive. Not clear from looking at the UI.
You shouldn't need to create an account, this is a bug and we've got a fix rolling out. Sorry you hit that!
Looks good, will try it out. All the best.
[dead]
“Analyzes email as you read them”
I have 108k unread emails on Gmail. “Analysis” as I read them isn’t my pain-point. Surfacing the 2-3 important emails I get daily so that I don’t miss something important among all the noise is really all I want from AI.
Everyone is focused on helping me write emails or summarize emails; things I frankly do not care about.
That said, I like the idea of a copilot for the browser, it just requires a great deal of trust in your company since I’m giving you access to my entire life.
I use a bunch of AI tools, I pay for Claude and OpenAI, I use Claude Code/Aider, and I've been playing with Gemini and Aider recently. Spending money on AI/LLMs is not foreign to me at all. That said:
The price on this is hilarious. You want me to pay another $20/mo for access and a fun new abstraction of "Message Credits". No problem, I'll just scroll down to the FAQ to figure out what a message credit is... nope, no definition at all.
There is little chance I'm going to pay $20/mo for a wrapper that is just 1 step over copy/paste [0] and absolutely no chance I'm going to pay for it with limits that aren't even explained.
[0] I'm sure this will be debated, but I'm not seeing anything special here.