Great idea! I've been humorously referring to chat agents as next gen Clippy because of their chipper, talky default personas which I find insufferably annoying.
I'm kind of shocked Microsoft didn't already do this as an alt version of their CoPilot UI. Really a huge miss on their part because I hate the overbearingly intrusive way they keep forcing it into their OS, apps and my fucking laptop keyboard. If they at least acknowledged their behavior and owned it (with a sly wink), I'd hate it a little less. I might even be up for a "Clippy is my CoPilot" sticker on my laptop (calling back to the old 80s "Jesus is my Copilot" bumper stickers).
> I'm kind of shocked Microsoft didn't already do this as an alt version of their CoPilot UI.
Seriously! This makes me think nobody at Microsoft with the authority to approve something like that has a sense of humor and/or good business sense. The nostalgia would be enormous. Hell I'm a linux person now and I'd install Clippy if it supported Fedora
Clippy was a laughing stock and target of derisive comedy for years. It has such bad brand recognition that nobody should be surprised that they aren’t using it.
But attaching a Clippy to a language model? Still nominally useless, but mindfully so!
It would be self-deprecating (un-deprecated???) humor for Microsoft, which would take the edge off of the often pushy and tone-deaf corporate look they continually and crassly paint themselves into by default.
And actually potentially useful as a branding touchstone: a visual and interface link across otherwise seemingly disparate model interfaces. Clearly delineating and bridging MS AI tools from all the other mixes of tools we are accumulating.
They could lean into the “clip” in Clippy with a side app for saving and organizing clippings and logs of notable interactions with any MS model, akin to a notes app. With features for compressing convos into compact topic cheat sheets (with retained sources & convos), lists and other helpful info gathering and leveraging tasks.
An ongoing accumulated compressed common core of context for both (hu)man and machine, er … Clippy.
Do it on April 1. With a “we told you so” tagline. Have a few 2025/6 templates, like writing a presidential executive order because there are so many of them.
clicked this link on mobile and every page scroll caused another malicious ad redirect (??) there's also a huge bouncing "remove ads" button with an X that opens an advert in the background. can't tell if the ads are on purpose or if the owners have just ticked every ad network box
You're not wrong, but both mullvad's free DNS base filter (here: https://github.com/mullvad/encrypted-dns-profiles ) and Wipr blocked it on my iphone. Android just use ublock origin with firefox or another variant.
That kind of sucks, because there's AI LLM's just about everywhere else now. Even those customer service "live chat" windows are typically AI first. What are Ask Jeeves doing?
I hope you accept that likening how it is intended, and I can't imagine that being a good thing. Clippy was universally panned. To me, I wouldn't be telling people that the thing I'm spending time working on was received as this generation's Clippy.
Actually this is a good way to find product ideas. I placed a query in Grok to find posts about what people want, similar to this. Then it performs multiple searches on X including embedding search, and suggested people want stuff like tamagotchi, ICQ etc. back.
IIRC correctly, Clippy’s most famous feature was interrupting you to offer advice. The advice was usually basic/useless/annoying, hence Clippy’s reputation, but a powerful LLM could actually make the original concept work. It would not be simply a chatbot that responds to text, but rather would observe your screen, understand it through a vision model, and give appropriate advice. Things like “did you know there’s an easier way to do what you’re doing”. I don’t think the necessary trust exists yet to do this using public LLM APIs, nor does the hardware to do it locally, but crack either of those and I could see ClipGPT being genuinely useful.
The way I remember it a lot of software had "help" documentation with full text search in the late 1980s and early 1990s but the common denominator was that it didn't work in the sense that you got useful answers less than 10% of the time. Until Google came along, users got trained to avoid full text search facilities.
The full text facility attached to Clippy really was helpful, getting useful answers around 50% of the time. I thought the whole point of making him an engaging cartoon character was to overcome the prejudice mid-1990s users had towards full-text search in help.
We are probably getting closer to that with the newer multimodal LLMs, but you'd almost need to take a screenshot on intervals fed directly to the LLM to provide a sort of chronological context to help it understand what the user is trying to do and gauge the users intentions.
As you say though, I don't know how many people would be comfortable having screenshots of their computer sent arbitrarily to a non-local LLM.
> As you say though, I don't know how many people would be comfortable having screenshots of their computer sent arbitrarily to a non-local LLM.
Of the technical, hang-out-on-HN crowd? Ya, probably not many.
Of the other 99.99% of computer users? The majority of them wouldn't even think about it, let alone care. To quote a phrase, ”the user is going to pick dancing pigs over security every time”.
Even without the non-chalent attitude towards security, the majority of the population has been so conditioned that everything they do on a computer is already being sent to 1) Apple, 2) Google, 3) Microsoft, or 4) their employer, that they're burnt-out of caring.
All that is to say that if you can make a widely-available real-time LLM assistant that appeals to non-technical users, please invite me to your private-island-celebrity-filled-yacht-parties.
I think we're well into the paradigm of "hidden employee activity monitoring software" already taking periodic screenshots and sending it to an LLM somewhere, which then generates aggregate performance metrics and dashboards for managers. I've heard of multiple companies working on this for $bigcorp environments, customer service/call center workstation PCs, etc.
I couldn't find how to get back to the normal chat screen from settings easily, and loading the same model file that works in LM studio crashed my computer.
Wow. The ease-of-use is insanely good. I haven't figured out yet how to move clippy to a different location on the screen (rather than centred), but it works well. I have multiple models downloaded and am chatting already!
Fun fact: the newest generation (such as myself, a 23 year old programmer) were actually not even alive when Clippy existed. I only know of it from an Office reference. One day I will have something like that -- maybe MSN or internet explorer?
Super cool. Serious 90s vibes. I also tried to make a super clippy here. https://chatbotkit.com/examples/super-clippy I think I match the color shema perfectly but does not have the same feeling as the original.
Funny. But you know, with multimodal models perhaps someone will finally crack when it is appropriate to interrupt someone with a relevant suggestion. I think I would like a personal assistant that would be able to say, "Hey, you have been debugging this one function for 5 hours and you still have 3 more to fix by EOB. Would it make sense to pause for a bit and see if other fixes could be done quickly?"
Really interesting project. I love the combination of LLM with a 90s aesthetic. Great that it works with a really simple configuration and runs offline
Looks like it's a special font provided by https://github.com/jdan/98.css (Which has come a long way in the past couple of years, despite still being 0.1.x)
I should probably 1.0 it and call it a day, it's pretty much done. In the back of my mind I've thought of making the markup more amenable to LLMs like Sonnet (maybe with tailwind style utility classes)
The real answer is that some of us (the Electron maintainers) have been playing with local LLMs in desktop apps and right now, node-llama-cpp is by far the easiest way to experiment - but it's also not meant for desktop apps and hence has _a lot_ of dependencies.
In general, pruning libraries in Electron isn't as easy as it should be - it's probably something for us to work on.
I think it’s legitimate to ask why these dependencies are necessary. LLMs have created whole new classes of vulnerabilities, and things like a GitHub client (which downloads arbitrary data/code) and a templating engine (which executes it) expose an even larger attack surface.
If someone’s going to get RCE on my machine, I don’t want it to be through the silly Clippy LLM UI, you know?
> A JavaScript implementation of the Jinja templating language
A guess without looking into the code: Jinja templating is used to define how to prompt the model (i.e. system first, then this specific character / token, then user, then if it's a tool prepend this and append that, etc.)
I think this is explained on the linked project page:
This project isn't trying to be your best chat bot. I'd like you to enjoy a weird mix of nostalgia for 1990s technology paired with one the most magical technologies we can run on our computers in 2025.
You might be looking for the more minimalist Grumpy which is hand-hewn from a pure silicon monocrystal.
Actually this says the trademark is pending them proving they are using it and they aren’t so they keep filing extensions. Also it’s limited in scope to word processing (at there is lots of back and forth with the trademark office about that)
Microsoft uses Clippy for the paperclip emoji in the Fluent Emoji set. The trademark is why the open source version of Fluent Emoji doesn't use Clippy's likeness.
Question for the devs in here...something I've been thinking about a lot recently. So I see that OP linked out to a public github repo...but when downloading the actual bundle, what's a quick way for me to determine that what I'm installing on my mac is actually the same as what's in the public repo? It's always seemed like a loophole to me ready for (potential) exploitation.
>> Ship project.
>> Link out Github repo on the static site somewhere
>> Gain trust instantly as users presume the public repo is what's used behind the scenes
Disclaimer: I'm a web dev and don't know a single thing about native MacOS software
I sign my binaries on macOS with Apple codesign and notarize - and with Microsoft's Azure trusted signing for Windows. Both operating systems will actually show you a lot of warning dialogs before running anything unsigned. It's far from perfect - but I do wish we'd get more into the habit of signing binaries, even if open source.
It was great / depressing to mention Clippy at a recent meetup and see the generational divide between those who groaned and everyone who looked confused.
You sure wouldn't want them spying on you, stealing your data, chewing up your resources for shady profit schemes, or making your machine unbootable. Better to leave that to the experts at Microsoft and FAANG since all those features come preinstalled nowadays.
Snark aside, given the context, this really seems like a baseless attack on independent open source developers, who represent a significant potion of this site's subject matter and target audience. Genuine question: why do you feel that this warning is appropriate here but not the dozens of other solo github projects that make it to the HN front page every week?
I'm all for these prepackaged local-only AI projects. Much more my speed than corporate cloud services. Real shame this one went down the path of choosing an embodiment that makes me want to shoot holes in my screen. It's even worse than those pixel art cats that chase my cursor on certain blogs. I miss plenty of things about the 90s, but I seriously doubt I'll live long enough to forget how much Clippy is not one of those things. Clippy would be more suitable for a horror game than an assistant. Going out of their way in the README to profusely thank Microsoft for summoning that hellspawn is just icing on the cake.
I hate to put down anyone's open source hobby project, and the guy looks so friendly and happy in his picture. But my honest reaction is fear of what further nightmares people are going to start animating with AI. I'd rather be hunted by a Boston Dynamics robot than have to face Clippy on my screen every day. Might as well add Rover from Microsoft Bob, some blink/marquee tags, a MIDI file playing in the background, and a minigame about diagnosing DMA conflicts in mixed plug and play and non-PnP systems. Some parts of the 90s should stay in the 90s.
> I'd rather be hunted by a Boston Dynamics robot than have to face Clippy on my screen every day.
This is the first AI thing I've actually bothered to install on my computer. Until today, despite being a technologist, I've only played with AIs via browser. I think AIs are interesting and can be useful but, having retired early, I'm not writing code or work emails so there hasn't been any compelling need.
I've thought about installing a local LLM to just play around with, but I have a long list of other things to play with (pinball machines, music making, photography, vintage video games) and AI just never got to the top of the list. I think I was also resistant because chat interfaces tend to be so annoying. I hate it when they LARP being a human. Giving a chat agent a retro 90s UX that's legendary for being annoying and clueless just seems so... on message, I thought "Yeah, I can probably not hate using this..."
I'm in the exact same boat as you describe, except it was some other precompiled local-only project instead of this one that I tried a few months ago. That's why I said I like projects like these, because it's a fully private hassle free way to try out LLMs. Haven't really figured out any good purposes for it in my life, but I like to see these tools being made available to people without the time/motivation/savvy to jump through a bunch of hoops.
The Clippy character specifically is the part I find off-putting, but perhaps that's just an excess of relevant experience. How many times I had to explain to confused people that it's not saying anything you have to care about, or disable it for them when they're cursing at their screen because the "hide" option doesn't actually disable it you have to go into the settings for that or it keeps popping up. Which made it just another config burden when I'd be installing office on many computers in a day.
Now, a strong argument could be made that those experiences have made me unreasonable and bigoted against animated paperclips, because this is not the original Clippy. I can live with that.
> Which made it just another config burden when I'd be installing office on many computers in a day.
Ah, well that does explain why you have some... baggage. My experience was different as I wasn't supporting or interfacing regularly back then with anyone who wasn't tech savvy. I endured Clippy for about 30 seconds, realized it was a stupid idea only a big corporation would think was cool or useful, turned it off and moved on with my day.
... but I think we may be heading for a new 'golden age' of web animation and gratuitous creativity. Personally, I'm happy to see more crazy animated stuff, it's the corporate dark patterns and bad UX that I hate.
You link your os to a local or cloud llm, and a local program asking the OS for a response and can’t even tell which one you’re using or whether it’s on the machine or not. It should all be abstracted away.
To me, the value of these types of projects is specifically that they are self-contained and local-only. That's the only kind of interaction with it I'm comfortable with right now. I mostly jumped ship on commercial software a long time ago, so I'm hoping there will still be some AI-free linux distros for a good long time. Different strokes for different folks, I suppose. At the point that the type of AI integration you're imagining becomes ubiquitous and mandatory, I may or may not stop working with computers entirely, depending on the state of the tech and the state of society by then.
There's a number of standard APIs already, OpenAI supports Anthopic's MCP, LM studio supports both their proprietary API as well as OpenAI's API. OpenAI has open sourced their realtime API (https://github.com/openai/openai-realtime-console/tree/webso...) and others. Most local clients just have a https://URL:port and then a drop down box for which RESTful API you want to use (for 88% of use cases they all support the same stuff, for realtime it's not quite settled yet), plus a field for an API key if needed.
i'm not sure if this post was written with humor as intent, but i found it hilarious. ive never heard someone talk about clippy with such disdain.
> I'd rather be hunted by a Boston Dynamics robot than have to face Clippy on my screen every day.
this is something else. i dealt with clippy when i was younger but i only have fond memories. it was useless, but it brought personality to an otherwise fairly mundane product.
I'm glad! :) I do actually feel some less exaggerated version of what I wrote, but the excess in the verbiage was largely comedic. If you look it up pretty much anywhere, you'll find that there's a very large camp of us Clippy haters who never recovered. I was doing some amount of IT support at the time, and one of the main problems was all the popping up and asking questions confused people in various ways, and if you hid it the obvious way it'd just pop up again. Back when computers were still a new and novel thing for many people, having constant offers of "help" popping up when you're just trying to type a letter introduced counterproductive amounts of cognitive load for some frustrated users I got to deal with.
and...? it does not change the fact that this "app" would get 0 attention if it wasn't using nostalgic IP that does not belong to the developer. their are undoubtedly better, more original apps being posted to HN right now that likely deserve the attention more, but they're not using stolen IP to get attention, so they don't.
Great idea! I've been humorously referring to chat agents as next gen Clippy because of their chipper, talky default personas which I find insufferably annoying.
I'm kind of shocked Microsoft didn't already do this as an alt version of their CoPilot UI. Really a huge miss on their part because I hate the overbearingly intrusive way they keep forcing it into their OS, apps and my fucking laptop keyboard. If they at least acknowledged their behavior and owned it (with a sly wink), I'd hate it a little less. I might even be up for a "Clippy is my CoPilot" sticker on my laptop (calling back to the old 80s "Jesus is my Copilot" bumper stickers).
> I'm kind of shocked Microsoft didn't already do this as an alt version of their CoPilot UI.
Seriously! This makes me think nobody at Microsoft with the authority to approve something like that has a sense of humor and/or good business sense. The nostalgia would be enormous. Hell I'm a linux person now and I'd install Clippy if it supported Fedora
Clippy was a laughing stock and target of derisive comedy for years. It has such bad brand recognition that nobody should be surprised that they aren’t using it.
I think that is what makes it a humor goldmine.
Clippy was useless.
But attaching a Clippy to a language model? Still nominally useless, but mindfully so!
It would be self-deprecating (un-deprecated???) humor for Microsoft, which would take the edge off of the often pushy and tone-deaf corporate look they continually and crassly paint themselves into by default.
And actually potentially useful as a branding touchstone: a visual and interface link across otherwise seemingly disparate model interfaces. Clearly delineating and bridging MS AI tools from all the other mixes of tools we are accumulating.
They could lean into the “clip” in Clippy with a side app for saving and organizing clippings and logs of notable interactions with any MS model, akin to a notes app. With features for compressing convos into compact topic cheat sheets (with retained sources & convos), lists and other helpful info gathering and leveraging tasks.
An ongoing accumulated compressed common core of context for both (hu)man and machine, er … Clippy.
Do it on April 1. With a “we told you so” tagline. Have a few 2025/6 templates, like writing a presidential executive order because there are so many of them.
It was, but as someone without a dog in the hunt, I loved that ol' Clipmeister.
Especially the old 'suicide note' joke image... guess would be called a meme today.
yeah, make it give edgy suggestions like: Do you want to find a new job?
They did though, I swear it was in a presentation you could select clippy as an avatar.
Edit: yes found it.
[1] https://windowsreport.com/with-copilot-avatar-microsoft-will...
clicked this link on mobile and every page scroll caused another malicious ad redirect (??) there's also a huge bouncing "remove ads" button with an X that opens an advert in the background. can't tell if the ads are on purpose or if the owners have just ticked every ad network box
You're not wrong, but both mullvad's free DNS base filter (here: https://github.com/mullvad/encrypted-dns-profiles ) and Wipr blocked it on my iphone. Android just use ublock origin with firefox or another variant.
Genuine question, if you're willing to indulge me: Why aren't you using ad-blocking of one type or another?
(Assumption: You're tech literate, given the audience of this website. So I tend to assume it must be a conscious decision not to use adblocking)
I don't browse without it these days.
> mobile
i have ublock origin on my pc and macbook. trying firefox mobile with ublock but it's still habit to open chrome on my phone
Strong recommend for Android, Firefox and uBlock Origin
I also have these Extensions:
Privacy Badger and UBlock Origin don't work well together, and LocalCDN is better than Decentraleyes
Sorry didn't notice, it was the top google result. I have ublock origin and firefox so I tend not to see many ads.
There's a lot of missed opportunities out there. For example, AskJeeves is still just a vanilla search engine (Google front-end).
That kind of sucks, because there's AI LLM's just about everywhere else now. Even those customer service "live chat" windows are typically AI first. What are Ask Jeeves doing?
Early on I gave it a custom instruction:
They will. It’s a no brainer to add a visual to the personality.
They can bring back clippy, Cortana, and all the other variants, in classic or modern mode. Hell why not a BonziBuddy knockoff.
An opportunity for Carmen Sandiego as well.
I'm firmly of the opinion that if they had shipped what is copilot as Cortana, they'd have seen little to no backlash.
Just had ChatGPT make this: https://ibb.co/pB4SPJBW
Am I right to feel wary of clicking this link? My spidy sense says 'don't do it'.
it's an ibb.co link; ibb.co is just an free online image host (the link lets you preview the uploaded image)
That’s a lot of hate you’re channeling there.
Customers I build AI chat features for also liken it to clippy. I think it’s a very common association.
I hope you accept that likening how it is intended, and I can't imagine that being a good thing. Clippy was universally panned. To me, I wouldn't be telling people that the thing I'm spending time working on was received as this generation's Clippy.
Clippy was panned because it was intrusive and offered very little real help, but the design and concept themselves were always popular.
you just described modern LLM bots as well
A nice addition (unless I missed it) would be to add an existing API key for remote model access?
This is a clear case of "Build Something People Want".
After all it was requested almost daily over at x.com
https://x.com/search?q=ai%20bring%20clippy%20back&src=typed_...
Actually this is a good way to find product ideas. I placed a query in Grok to find posts about what people want, similar to this. Then it performs multiple searches on X including embedding search, and suggested people want stuff like tamagotchi, ICQ etc. back.
Fun fact: clippy came from Microsoft Bob, which Melinda Gates was the marketing manager for.
I have often wondered what role their relationship played in keeping Clippy around. And now I wonder if Clippy makes Bill Gates sad since the divorce.
>And now I wonder if Clippy makes Bill Gates sad since the divorce
I doubt he thinks about clippy much at all.
> I doubt he thinks about clippy much at all
Guys I think I found Bill’s HN handle
Can you add narration in Gilbert Gottfried's voice?
https://www.youtube.com/watch?v=tu_Pzuwy-JY
IIRC correctly, Clippy’s most famous feature was interrupting you to offer advice. The advice was usually basic/useless/annoying, hence Clippy’s reputation, but a powerful LLM could actually make the original concept work. It would not be simply a chatbot that responds to text, but rather would observe your screen, understand it through a vision model, and give appropriate advice. Things like “did you know there’s an easier way to do what you’re doing”. I don’t think the necessary trust exists yet to do this using public LLM APIs, nor does the hardware to do it locally, but crack either of those and I could see ClipGPT being genuinely useful.
The way I remember it a lot of software had "help" documentation with full text search in the late 1980s and early 1990s but the common denominator was that it didn't work in the sense that you got useful answers less than 10% of the time. Until Google came along, users got trained to avoid full text search facilities.
The full text facility attached to Clippy really was helpful, getting useful answers around 50% of the time. I thought the whole point of making him an engaging cartoon character was to overcome the prejudice mid-1990s users had towards full-text search in help.
It looks like you're writing a letter.
Would you like help?
* Get help with writing the letter
* Just type the letter without help
|_| Don't show me this tip again
>and give appropriate advice
"It's time to work, Dave"
I’m sorry, I can’t do that Hal
Even funnier would be to make it unnecessarily mean and vexing.
Wait, are you really looking this up? You don't even know how to do this? Are you kidding me?
Gosh, it's been an hour and you still haven't fixed this bug? Are you retarded or something? You don't deserve this job.
I already have a little voice in my head that tells me those things!
That said, if we could automate it, it might free up more of my brain for productivity…
We are probably getting closer to that with the newer multimodal LLMs, but you'd almost need to take a screenshot on intervals fed directly to the LLM to provide a sort of chronological context to help it understand what the user is trying to do and gauge the users intentions.
As you say though, I don't know how many people would be comfortable having screenshots of their computer sent arbitrarily to a non-local LLM.
> As you say though, I don't know how many people would be comfortable having screenshots of their computer sent arbitrarily to a non-local LLM.
Of the technical, hang-out-on-HN crowd? Ya, probably not many.
Of the other 99.99% of computer users? The majority of them wouldn't even think about it, let alone care. To quote a phrase, ”the user is going to pick dancing pigs over security every time”.
Even without the non-chalent attitude towards security, the majority of the population has been so conditioned that everything they do on a computer is already being sent to 1) Apple, 2) Google, 3) Microsoft, or 4) their employer, that they're burnt-out of caring.
All that is to say that if you can make a widely-available real-time LLM assistant that appeals to non-technical users, please invite me to your private-island-celebrity-filled-yacht-parties.
> I don't know how many people would be comfortable having screenshots of their computer sent arbitrarily to a non-local LLM
shudders.
So, the Replay feature being slowly rolled out in Win11?
I think we're well into the paradigm of "hidden employee activity monitoring software" already taking periodic screenshots and sending it to an LLM somewhere, which then generates aggregate performance metrics and dashboards for managers. I've heard of multiple companies working on this for $bigcorp environments, customer service/call center workstation PCs, etc.
Microsoft infamously is adding AI to Windows to constantly watch your screen and people understandably are not super excited for it.
I personally can’t wait to ask to recall something I saw before but can’t quite remember where.
Pretty soon I won’t even need biological memory.
if it ran entirely on the local machine and didn't send information back to Microsoft I think people would be far more accepting of it.
It can still be annoying; I feel it is part of his personality.
It looks like you are writing a comment on Hacker News.
Would you like help with:
- Commas? There shouldn't be one behind "responds to text"
- Capitalization? You've missed a D in "did you know..."
- Punctuation? You've missed a question mark behind "what you’re doing". It goes inside the quotes, of course!
[] Don't ever suggest anything like this ever again.
Amazing, we also have CowPilot now https://github.com/agentsea/cowpilot
I couldn't find how to get back to the normal chat screen from settings easily, and loading the same model file that works in LM studio crashed my computer.
I like the idea, though.
Great job! Having ollama support would be useful as well[1]! [1]https://github.com/ollama/ollama
Wow. The ease-of-use is insanely good. I haven't figured out yet how to move clippy to a different location on the screen (rather than centred), but it works well. I have multiple models downloaded and am chatting already!
click and drag on the bottom right corner of clippy to drag
Can this keep popping up, interrupt you, and have the most annoying voice ever added please?
Man do I ever miss this UI design. Nice work!
Pretty fantastic follow-up to https://www.latent.space/p/clippy-v-anton
One of my very first AI projects was in the late 90s and used the Microsoft Agent API (which Clippy uses) as the interface.
It used Merlin rather than Clippy and was extremely basic as AI. But it was a fun project.
I recently did this in our main system that we recently added an LLM feature to (for fun internally, not sending to prod) using
https://github.com/pi0/clippyjs
Finally a useful UI for llama.cpp!
Thank you Felix! This is extremely cool! Can you please make a short blog post explaining how is it technically implemented?
Love it.
On macOS it always launches in the middle of the screen - is there a way to move it around?
To move clippy you want to drag the piece of paper on which clippy sits -- clicking clippy himself will hide and show the chat window.
Fun fact: the newest generation (such as myself, a 23 year old programmer) were actually not even alive when Clippy existed. I only know of it from an Office reference. One day I will have something like that -- maybe MSN or internet explorer?
It's not quite that bad! The last version of Office released with Clippy was in May 2003! So you would have been born, if only just.
Yeah I was a tiny little two year old:) Definitely wasn't coding yet hahaha
Super cool. Serious 90s vibes. I also tried to make a super clippy here. https://chatbotkit.com/examples/super-clippy I think I match the color shema perfectly but does not have the same feeling as the original.
It's way too high resolution!
Animations are missing too.
When do we get the BonziBuddy reskin?
I still haven't gotten over the trauma of Clippy 1.0.
Man, brings back memories I didnt even think I still hadClippy was kinda ridiculous back then but Id 100% mess with this now tbh.
It looks like you're talking about a cartoon assistant character. Would you like help?
ICYDN: The proper name of Clippy is actually "Clippit", as introduced in Office 97.
Feedback: I think it would be very helpful for users to know ahead of time what kinda performance they can expect based on their system.
Will this properly interrupt me in the middle of flow and ask unrelated questions, or is it just another clippy knock-off?
Funny. But you know, with multimodal models perhaps someone will finally crack when it is appropriate to interrupt someone with a relevant suggestion. I think I would like a personal assistant that would be able to say, "Hey, you have been debugging this one function for 5 hours and you still have 3 more to fix by EOB. Would it make sense to pause for a bit and see if other fixes could be done quickly?"
Great idea and design, thanks for this! I was hoping since some time to see this :-D
I hope that one day a non-Electron app (to minimize resource usage when idle) will also appear!
It's weird that when clippy was new I found it to be everything that's wrong with UI design, and today I'm nostalgic for it
Nah, it's not weird. You said it yourself: nostalgia. It's human nature to romanticise the past. I bet you would hate it again if you used it today.
Pack it up boys, they finally made the killer app.
I can't get this to work on my aging 2017 Intel work mac.
> Error: Error invoking remote method 'ELECTRON_LLM_CREATE': Error: Error: NoBinaryFoundError
Ah, shoot, I had an error in my build script. That's now fixed!
https://github.com/felixrieseberg/clippy/releases/tag/v0.4.1
Is it insane that I tried to make a version of this exactly a week ago!? This is freakin awesome, congratulations!
Really interesting project. I love the combination of LLM with a 90s aesthetic. Great that it works with a really simple configuration and runs offline
Makes me think of this short story.
https://gwern.net/fiction/clippy
Any love for the other avatars? Power Pup? I think there were a few… Otherwise, thanks, this is great.
I love the terrible font rendering! Is it a special font, or some CSS?
Looks like it's a special font provided by https://github.com/jdan/98.css (Which has come a long way in the past couple of years, despite still being 0.1.x)
Although there is a CSS rule for manipulating how fonts are anti-aliased, it was never standardized, and Firefox doesn't implement the vital no-smoothing option: https://developer.mozilla.org/en-US/docs/Web/CSS/font-smooth
Maybe with enough retro revivals it will receive attention.
I should probably 1.0 it and call it a day, it's pretty much done. In the back of my mind I've thought of making the markup more amenable to LLMs like Sonnet (maybe with tailwind style utility classes)
Accept my deepest gratitudes for creating this functional art. Love the idea and execution and can't wait to use it!
This is such an amazing piece of work — truly impressive! Hats off to you If it supports Ollama and local LLMs too, it'll be absolutely unbeatable!
I love your website design.
They definitely missed on using underlines for headings
Like phoenix, it rises from the ashes..
This is a thing of beauty, thank you!! :-D
BonziBuddy next?
Followed by Tiny Elvis?
https://somethingorotherwhatever.com/tiny-elvis/
Revenge of the paperclips
Clippy was peak Windows. Everything went downhill since...
Awesome! Now I just want Perplexity to acquire the AskJeeves brand.
I thought Clippy first shipped in XP
Nope, I remember it in Office 97. Which was released in 1996, of course.
Next up, "Rover" the dog from Microsoft Bob
https://fabulous.systems/posts/2024/06/if-i-ever-get-a-dog-i...
yes yes yes!!!
"I can fix him"
This is cool, but does no one even look at what libraries they're shipping anymore? I mean, why does this Clippy-style LLM interface bundle:
- A JavaScript implementation of the Jinja templating language
- A full GitHub API client
- A library that takes a string and tells you if it's a valid npm package name
- A useless shim for the JavaScript Math module
And 119 other libraries? This thing would have taken up 10% of the maximum disk space available on a Windows 95 FAT16 volume.
The real answer is that some of us (the Electron maintainers) have been playing with local LLMs in desktop apps and right now, node-llama-cpp is by far the easiest way to experiment - but it's also not meant for desktop apps and hence has _a lot_ of dependencies.
In general, pruning libraries in Electron isn't as easy as it should be - it's probably something for us to work on.
So to be clear, your complaint is that the nostalgia Clippy app that puts a cartoon paper clip on your desktop, isn't efficient enough?
I think it’s legitimate to ask why these dependencies are necessary. LLMs have created whole new classes of vulnerabilities, and things like a GitHub client (which downloads arbitrary data/code) and a templating engine (which executes it) expose an even larger attack surface.
If someone’s going to get RCE on my machine, I don’t want it to be through the silly Clippy LLM UI, you know?
Maybe it was vibe coded and the libraries were added while going down paths that turned out to be dead ends and the LLM never cleaned up after itself?
People have been perfectly capable of making that mistake themselves since long before "vibe coding" existed.
> A JavaScript implementation of the Jinja templating language
A guess without looking into the code: Jinja templating is used to define how to prompt the model (i.e. system first, then this specific character / token, then user, then if it's a tool prepend this and append that, etc.)
It seems that this is possibly not necessary, since LLaMA.cpp already integrates Jinja with CPP implementation (through minja)
I think this is explained on the linked project page:
This project isn't trying to be your best chat bot. I'd like you to enjoy a weird mix of nostalgia for 1990s technology paired with one the most magical technologies we can run on our computers in 2025.
You might be looking for the more minimalist Grumpy which is hand-hewn from a pure silicon monocrystal.
Do this one next: https://en.wikipedia.org/wiki/BonziBuddy
Pretty sure Clippy is trademarked. Had the same idea but did not go through with it because of the TM.
Correct.
https://tsdr.uspto.gov/#caseNumber=90782096&caseSearchType=U...
Actually this says the trademark is pending them proving they are using it and they aren’t so they keep filing extensions. Also it’s limited in scope to word processing (at there is lots of back and forth with the trademark office about that)
The character is actually named Clippit. Although maybe MS trademarked Clippy after it became the more common name.
I seriously doubt Microsoft would enforce it for a non-commercial side project.
Trademarks have to actually be used to remain enforceable, I think. Not sure MS could claim Clippy after all this time, not that they might not try.
Three fact that it's not a product anymore doesn't mean it's unused; a quick search says they at a minimum used it in 2021: https://www.microsoft.com/en-us/microsoft-365/blog/2021/07/0...
Wow, had no idea. Like a bad penny.
Microsoft uses Clippy for the paperclip emoji in the Fluent Emoji set. The trademark is why the open source version of Fluent Emoji doesn't use Clippy's likeness.
Question for the devs in here...something I've been thinking about a lot recently. So I see that OP linked out to a public github repo...but when downloading the actual bundle, what's a quick way for me to determine that what I'm installing on my mac is actually the same as what's in the public repo? It's always seemed like a loophole to me ready for (potential) exploitation.
>> Ship project. >> Link out Github repo on the static site somewhere >> Gain trust instantly as users presume the public repo is what's used behind the scenes
Disclaimer: I'm a web dev and don't know a single thing about native MacOS software
Yeah, reproducible builds would be fantastic.
I sign my binaries on macOS with Apple codesign and notarize - and with Microsoft's Azure trusted signing for Windows. Both operating systems will actually show you a lot of warning dialogs before running anything unsigned. It's far from perfect - but I do wish we'd get more into the habit of signing binaries, even if open source.
you don't, that is what reproducible builds are trying to solve, but even then it would still need someone to compile and check.
https://en.wikipedia.org/wiki/Reproducible_builds
It was great / depressing to mention Clippy at a recent meetup and see the generational divide between those who groaned and everyone who looked confused.
I’m waiting for BonziBuddy AI
[dead]
"...they didn't stop to think if they should.”
[flagged]
[flagged]
I hope people realize that this is an easy way to get a virus.
Don't install third party software except from highly trusted sources.
You sure wouldn't want them spying on you, stealing your data, chewing up your resources for shady profit schemes, or making your machine unbootable. Better to leave that to the experts at Microsoft and FAANG since all those features come preinstalled nowadays.
Snark aside, given the context, this really seems like a baseless attack on independent open source developers, who represent a significant potion of this site's subject matter and target audience. Genuine question: why do you feel that this warning is appropriate here but not the dozens of other solo github projects that make it to the HN front page every week?
Microsoft Defender didn't find anything
But BonzaiBuddy is your friend.
[flagged]
[flagged]
I'm all for these prepackaged local-only AI projects. Much more my speed than corporate cloud services. Real shame this one went down the path of choosing an embodiment that makes me want to shoot holes in my screen. It's even worse than those pixel art cats that chase my cursor on certain blogs. I miss plenty of things about the 90s, but I seriously doubt I'll live long enough to forget how much Clippy is not one of those things. Clippy would be more suitable for a horror game than an assistant. Going out of their way in the README to profusely thank Microsoft for summoning that hellspawn is just icing on the cake.
I hate to put down anyone's open source hobby project, and the guy looks so friendly and happy in his picture. But my honest reaction is fear of what further nightmares people are going to start animating with AI. I'd rather be hunted by a Boston Dynamics robot than have to face Clippy on my screen every day. Might as well add Rover from Microsoft Bob, some blink/marquee tags, a MIDI file playing in the background, and a minigame about diagnosing DMA conflicts in mixed plug and play and non-PnP systems. Some parts of the 90s should stay in the 90s.
> I'd rather be hunted by a Boston Dynamics robot than have to face Clippy on my screen every day.
This is the first AI thing I've actually bothered to install on my computer. Until today, despite being a technologist, I've only played with AIs via browser. I think AIs are interesting and can be useful but, having retired early, I'm not writing code or work emails so there hasn't been any compelling need.
I've thought about installing a local LLM to just play around with, but I have a long list of other things to play with (pinball machines, music making, photography, vintage video games) and AI just never got to the top of the list. I think I was also resistant because chat interfaces tend to be so annoying. I hate it when they LARP being a human. Giving a chat agent a retro 90s UX that's legendary for being annoying and clueless just seems so... on message, I thought "Yeah, I can probably not hate using this..."
I'm in the exact same boat as you describe, except it was some other precompiled local-only project instead of this one that I tried a few months ago. That's why I said I like projects like these, because it's a fully private hassle free way to try out LLMs. Haven't really figured out any good purposes for it in my life, but I like to see these tools being made available to people without the time/motivation/savvy to jump through a bunch of hoops.
The Clippy character specifically is the part I find off-putting, but perhaps that's just an excess of relevant experience. How many times I had to explain to confused people that it's not saying anything you have to care about, or disable it for them when they're cursing at their screen because the "hide" option doesn't actually disable it you have to go into the settings for that or it keeps popping up. Which made it just another config burden when I'd be installing office on many computers in a day.
Now, a strong argument could be made that those experiences have made me unreasonable and bigoted against animated paperclips, because this is not the original Clippy. I can live with that.
> Which made it just another config burden when I'd be installing office on many computers in a day.
Ah, well that does explain why you have some... baggage. My experience was different as I wasn't supporting or interfacing regularly back then with anyone who wasn't tech savvy. I endured Clippy for about 30 seconds, realized it was a stupid idea only a big corporation would think was cool or useful, turned it off and moved on with my day.
Eloquently put.
... but I think we may be heading for a new 'golden age' of web animation and gratuitous creativity. Personally, I'm happy to see more crazy animated stuff, it's the corporate dark patterns and bad UX that I hate.
I’d prefer it be an OS API.
You link your os to a local or cloud llm, and a local program asking the OS for a response and can’t even tell which one you’re using or whether it’s on the machine or not. It should all be abstracted away.
To me, the value of these types of projects is specifically that they are self-contained and local-only. That's the only kind of interaction with it I'm comfortable with right now. I mostly jumped ship on commercial software a long time ago, so I'm hoping there will still be some AI-free linux distros for a good long time. Different strokes for different folks, I suppose. At the point that the type of AI integration you're imagining becomes ubiquitous and mandatory, I may or may not stop working with computers entirely, depending on the state of the tech and the state of society by then.
There's a number of standard APIs already, OpenAI supports Anthopic's MCP, LM studio supports both their proprietary API as well as OpenAI's API. OpenAI has open sourced their realtime API (https://github.com/openai/openai-realtime-console/tree/webso...) and others. Most local clients just have a https://URL:port and then a drop down box for which RESTful API you want to use (for 88% of use cases they all support the same stuff, for realtime it's not quite settled yet), plus a field for an API key if needed.
is it possible you're not the target audience?
Which part of my original comment made that a question worth asking? Thought I had already expressed that fairly clearly.
i'm not sure if this post was written with humor as intent, but i found it hilarious. ive never heard someone talk about clippy with such disdain.
> I'd rather be hunted by a Boston Dynamics robot than have to face Clippy on my screen every day.
this is something else. i dealt with clippy when i was younger but i only have fond memories. it was useless, but it brought personality to an otherwise fairly mundane product.
I'm glad! :) I do actually feel some less exaggerated version of what I wrote, but the excess in the verbiage was largely comedic. If you look it up pretty much anywhere, you'll find that there's a very large camp of us Clippy haters who never recovered. I was doing some amount of IT support at the time, and one of the main problems was all the popping up and asking questions confused people in various ways, and if you hid it the obvious way it'd just pop up again. Back when computers were still a new and novel thing for many people, having constant offers of "help" popping up when you're just trying to type a letter introduced counterproductive amounts of cognitive load for some frustrated users I got to deal with.
a very basic app getting a bunch of undue attention thanks to nostalgia for someone else's IP, classic
You almost sound bitter about it
and...? it does not change the fact that this "app" would get 0 attention if it wasn't using nostalgic IP that does not belong to the developer. their are undoubtedly better, more original apps being posted to HN right now that likely deserve the attention more, but they're not using stolen IP to get attention, so they don't.