Spaces:
Running
[ANNOUNCEMENT] π£ HuggingChat is closing for now
I have bittersweet news to share. π’
We are closing down huggingchat for now.
Huggingchat was launched in April 2023 when ChatGPT was just 5 months old and it was still quite hard to deploy a similar service. The goal of this app was to showcase that it was now possible to build an open source alternative to ChatGPT, building on top of the community's models and building blocks. πͺ
We also built Huggingchat as a testbed to optimize inference which led to the creation of text-generation-inference which in turn led to HF contributions to SGLang, vLLM, llama.cpp, and other LLM runtime engines since then.
In the past 27 months we've launched:
- π§ Assistants, an easy way to bundle a prompt, a model, and some tools into an easily-shareable URL you can send your friends
- π οΈ HuggingChat Tools, exposing the full ecosystem of Gradio Spaces straight from within the chat interface
- π¬ Many model launches were supported on HuggingChat: from OpenAssistant, all the LLamas, Phi-4, to Qwen, DeepSeek, Gemma.... π₯
The app has always been free and experimental. Today we are closing it to make room for something new and more integrated with the HF ecosystem. You can export your past conversations as a Zip file, to load them into another interface.
If you are looking for a cool alternative interface you should take a look at:
- LibreChat a clean and simple yet feature-packed chat interface
- openwebui β does it even need an intro anymore?
- Scira MCP, an up and coming MCP-centric chat interface
Finally, chat-ui (the codebase powering HuggingChat) is still maintained (GitHub repo).
You can 1-click deploy your own instance using the Chat UI Spaces Docker template.
What's happening to huggingchat π’ ?
I hope it's not too long.
no π
Also, how do I export my past conversations?
Hi everyone! π
This is also my last week at Hugging Face and I'm currently working on a simple page to let you export your personal data & conversation history from HuggingChat so you can reuse it on other platforms in the future. Let me know if you have any questions, I'm always happy to answer them.
It was a pleasure building this app for y'all π«‘
Thats sad.
I wished this never happened
ΠΠΠ’ Π’ΠΠΠ¬ΠΠ ΠΠ ΠΠ’Π!!!!! ΠΠ’ΠΠ ΠΠΠ’Π HUGGINGCHAT!
HuggingChat was truly superior to ChatGPT in many ways. Better tools, better models (GPT 4 actually kind of sucks), better everything. I totally understand that running something that powerful for free isn't sustainable, both financially and environmentally. Still sucks though.
Export feature should be up! Let me know if it works well for you.
It works. Thank you for that!
Sad to see the feature go
WTF?! This was my daily driver!
why is this happening? so sad
Hi everyone! π
This is also my last week at Hugging Face and I'm currently working on a simple page to let you export your personal data & conversation history from HuggingChat so you can reuse it on other platforms in the future. Let me know if you have any questions, I'm always happy to answer them.
It was a pleasure building this app for y'all π«‘
grateful. beyond gratitude. this was such a gift. I feel a sharp pain knowing this era is over. it made me believe and commit to text ai in a way I would not have otherwise pursued. I learned so much and felt guided by a vision bigger than me and OpenAI.
edit: my final remark should not be a jab at OpenAI, but it feels warranted. you gave us a feeling of invincibility during a time when the push to closed and tiered ai access felt inevitable. the community needs to rally here and anyone that depended on hfchat like me, we get it. don't let go of whatever that "it" feels like to you. we all glimpsed something huge when this was a beacon of hope in the storm
Oh, that's a great pity! π I've worked with this tool so often and was so happy to be able to use a free AI alternative based on open source. I was always amazed at all the things that could be done with this system. π
It goes without saying that such a system cannot be operated free of charge in the long term.
Thank you very much for your work and your time in maintaining this service for us. π€
Will HuggingChat be available for a fee in the future? As a kind of hosting service? That would be like many open source teams do: Develop an OpenSoucre product, those who want to can run it themselves on their own server and help where they can, and those who don't want to/can't run it can rent the service from the team for a fee.
Thank you for giving us the heads up, I didn't use this chat too much but, it had great personality or aionality. I can't wait to see what you are working on next. Also thank you for keeping the repo open, I will look into running this locally
Export feature should be up! Let me know if it works well for you.
what? you're not going to make us Upgrade to premium for the privilege of our own data? what kind of free service rug-pull is this? Where's the #ad sponsored HF-HugBox Brain Pills we were promised?
real shame, I used this all the time. Will something new be coming out soonish, and would it be similar to huggingchat?
Rest well, HC. π―οΈ You was incredibly useful for me as long as you was being.
β§α΄Ώα΄΅α΄Ύβ«βββ ββ(Γ³ Μ― Γ², )
upd:(I know it's not the end, but it's almost certain that nothing will be the same as before)
Wow. It was good, I suspected too good to be true. Hope it comes back in some form.
A project I was working on relied on HuggingChat :(
- What's the timeline on the alternative?
- Will the alternative be free for use as HuggingChat?
- Will it also have API support?
Are there any other free web-accessible ones anyone here recommends, particularly any that uses the Cohere model since it seems to have been the best of the uncensored ones.
why closed now
I felt it. I felt that something bad will happen. I was with HuggingChat since November/December 2023. It grew in front of my eyes. With HuggingChat, I saw AI advancing! I loved it. It helped me a lot through my life. I remember how assistants were released. Some time after, model parameters were made configurable. I spent nights having fun with models, maxing out all parameters to make them speak nonsense. Then, tools were released. I wanted to make my own alternative of default web search, as this feature was very outdated. Then, I discovered that you can't use your own tools in assistants. Some time after, I gave up on it. At this time, I noticed that the development is slowing down. There were some model updates, but they weren't impressive. I hoped that things will get better. But months passed, and now I see this message.
RIP HuggingChat.
DeepSeek killed HuggingChat!
I guess this is not the proper place to talk about that, but in despair...
My dream have always been to pay for HuggingChat (being able to pay in advance and consume my credit).
It has always been impossible.
I am not technically inclined enough, I guess, but maybe someone could help me to understand how to achieve this goal.
I mean, HuggingChat was not perfect, but its functionalities and interface were exactly what I was looking for, and they fit extremely well into my use cases.
How and where can I find the same (and pay for it in advance)?
The ability to switch models, models that can make search on internet/access url, upload attached documents to the requests, create my assistant*, and eventually use others, etc.,
is what I need.
*aka, profiles on which you can switch models
one of the best website on the internet ,free chat with Mutiple model i really miss HC hope it comes back
Will HuggingChat be available for a fee in the future? As a kind of hosting service? That would be like many open source teams do: Develop an OpenSoucre product, those who want to can run it themselves on their own server and help where they can, and those who don't want to/can't run it can rent the service from the team for a fee.
Yes! You can run your own HuggingChat instance on a Space here on HF (link to do so in the original post, it's easy to do - give it a try!).
Damn I almost made my mind reading chat app πππ
Are there any other free web-accessible ones anyone here recommends, particularly any that uses the Cohere model since it seems to have been the best of the uncensored ones.
Sorry, this is totally censored. We have a controversial figure in France named DieudonnΓ© MβBala MβBala whoβs been on the scene for three decadesβappearing in headlines, TV interviews, and talk shows thousands of timesβyet Cohere knows nothing about him, just as DeepSeek in China ignores Tiananmen. Please take those figures into account: decades of cumulative coverage on TV, in newspapers, internet and on radio. This isnβt a mistake or a coincidence*, but active censorship.
( And please, to all the trolls around, the point is not about this guy and if it is a good guy or a bad guy, the point is: it is a PROOF of an active censorship )
*Erasing a decadeβs worth of βmemoryβ about a wellβrepresented individual in an LLM is technically challenging and expensive: the most reliable approachβremoving all related training data and fully retraining (or at least reβfineβtuning) the modelβdemands thousands of GPU hours, extensive dataβengineering to prune direct and indirect references, and rigorous auditing to avoid leakage
π
Sad to hear!
Will you transfer the Github repo to somebody else?
That's just about the saddest thing that could have happened in the world of AI text generation.
Today we are closing it to make room for something new and more integrated with the HF ecosystem.
You mean a new text generation interface with languages to choose between? And hopefully soon?
We are cooking something new, please stay tuned...
Will it be also a llm chat?
Aww fuck no!
when it open
Aww fuck no!
damn we need open
Whatever comes next, please keep it free. I understand it can't be cheap running something like this, but it's such a service to the world to have this kind of resource available to everyone regardless of their ability to pay for it.
Hello - very sad - how will you communicate the finishing of the new cooking ?
And if i open the downloaded zip there are 3 files one firefox
assistance.json - nothing in
conversations.json - just the titles
file-675047856c70bad12a30e52b-faa8eecb - just one file with a part of only one conversation
Please how to get all the important chats?
Best wishes to all of you!
I came up with a great system prompt for a business assistant with proactive behavior this morning, after which I realized that I could finally do something. Now I have no way to continue this, because I used the Cover neural network, and the official website does not allow system prompts π₯Ή... Smaller affordable models simply can't handle the self-awareness and initiative mode, and even if they do, their capabilities are not sufficient for anything worthwhile. I canβt assemble a team of programmers because no one believes that conscious AI is even possible, so I had to go towards this dream alone with the AI. Today I finally got around to starting, and as always happens in life, something happened. Well...
the site does say "cooking something new, stay tuned", so I'd assume it's probably a replacement and it won't take too long (I hope) because there are really no good alternatives out there that are as simple and easy to use as Huggingchat.
Speaking of the alternatives not being easy to use, I've only ever used HuggingChat for these. I'm currently trying to set up LibreChat. What is key and how do I get one?
info on how to use LibreChat would be great, yeah.
So sad and so bad πΏπ±ππ±ππππ
Mmm
.
I loved this any alternatives
HF.co, thank you so much for taking the time and care to handle the data export.
Guys, where can we now find and explore all those amazing prompts (assistants) that you provided in HuggingChat?
Om what's happened
Well... Okay I guess
I'm assuming we'll be getting something new in a little while. the link to it is still in the side section on the huggingface site anyhow.
EDIT: the message that's replaced the Huggingchat page even says "we're cooking up something new, stay tuned". I know it's probably just wishful thinking, but even the announcement itself just says "for now".
@julien-c can you give us a little more direction? what timeframe are you looking at - days? weeks? months? Will it still be free?
Man i hope the assistants aren't lost forever they were amazing. thanks for the service guys hope they come back soon
HuggingChat was a rare gem in the LLM ecosystem. It was a space where things felt transparent, and customizable especially compared to the walled gardens of many corporate offerings.
I feeling a bit adrift right now especially for someone like me, who clearly thrives on the frontier of LLM experimentation
Have people considered using the Hugging Face inference playground? For people on the PRO plan everyone gets $2 in free inference credits so it can be used: https://huggingface.co/playground
Free plan gets a small quota as well
Hi, how do we ensure our data is deleted please? Without warning this service disappeared and now I can't find a way to delete my historical data.
I do not consent to my data being stored, used, sold or for any other use without my authorisation.
Will y'all guys atleast take a look at this ?
https://huggingface.co/spaces/huggingchat/chat-ui/discussions/682
This used to be my comfort zone, always providing me with something I need. When it was gone, I felt severely depressed for a few hours.
Lesson learned, comfort zones are a limited resource, and one shouldn't put all of them in one basket or on shaky, unreliable ground.
Setting up local llms doesn't seem like as bad of an investment anymore. At least they don't disappear when you don't expect it.
idk when it open guys
Hi, how do we ensure our data is deleted please? Without warning this service disappeared and now I can't find a way to delete my historical data.
I do not consent to my data being stored, used, sold or for any other use without my authorisation.
It was built from the begining to gather data so it will be a win win for us and them, so i think they're using our private data to train their models and for maybe advertising issues.
Hi, I think it makes sense to add a delete your data feature. We already had it before, it's just not exposed in the current page, let me bring it back up.
How do i open the conversations.json i exported librechat doesn't seem to support it
Letter accepting invitation to the UK
It is a pity that "helpers" from other users were not added to the export.
Hi, I think it makes sense to add a delete your data feature. We already had it before, it's just not exposed in the current page, let me bring it back up.
That will be very helpful, thank you
Hi, how do we ensure our data is deleted please? Without warning this service disappeared and now I can't find a way to delete my historical data.
I do not consent to my data being stored, used, sold or for any other use without my authorisation.It was built from the begining to gather data so it will be a win win for us and them, so i think they're using our private data to train their models and for maybe advertising issues.
Not according to the latest privacy policy, that's why I choose to use this service because it offered a level of privacy that others didn't. I hope that continues with the next offering.
I like Hugging Face chat a lot.
We could choose the proper model for different task, tune my own assistants to make them speak in the style we preferred, and integrate community tools to achieve versatile missions.
It's quite sad to hear the service has been closed so abruptly, without buffering period or announcement beforehand.
I'm an open source fan, rarely used propriety products, but lack of reliability is a drawback, just like HF chat's sudden shutdown this time.
I appreciate HuggingChat's efforts for the service, and feel sad about its disappearance.
PS Just a reminder for Hugging Face team,
Users may not put so much trust and confidence on your new products next time due to this experience.
After all, no one likes to risk their daily AI assistant or handy tools.
We have to search, compare, and read lots of stuff before dwelling into another platform.
what ?
is there any idea of what is the new features you will introduce ?
Please let there be an import feature in the hf chat-ui so that we may have a viewable experience again in the ui!
Ohh, I think I'm gonna go on a bender after this news.
HuggingChat was on a shortlist of chatbots recommended for my Mozilla Firefox sidebar, and after comparing it side by side with ChatGPT I found it to provide more valuable information with less redundancies and a more professional tone. Thanks for all the great work for all this time!
is it still available in a paid subscription?
The new chat ui like in that other post might be the reason?
Isn't this one of the main reasons for local servers' existence though, to make necessary changes and push them afterwards to streamline things and so that a sass doesn't suffer downtime? Both the announcement title and message are rather ambiguous and the "Discuss on the Hub" link seems redundant since developer feedback isn't there.
That said, I am in no position to complain since neither do i own the platform nor am I involved in development, but "We are cooking something new, please stay tuned..." could mean "it's been nice all these years but we've decided to shut the Hugging Chat service down." or "We'll be back after the changes, but don't worry, you'll still get the Hugging Chat that you loved so much, only better".
Today we are closing it to make room for something new and more integrated with the HF ecosystem.
Company speak for "it's dead and never coming back again".
Is there any sort of alternative we can go to in the meantime? Preferably free and hopefully uncensored
Is there any sort of alternative we can go to in the meantime? Preferably free and hopefully uncensored
Except running it locally, which I'm sure many of us are not in a position to do, from what I've seen. No.
LibreChat requires API keys, which means paying.
Well, RIP. Was a real good run though.
HuggingChat was on a shortlist of chatbots recommended for my Mozilla Firefox sidebar, and after comparing it side by side with ChatGPT I found it to provide more valuable information with less redundancies and a more professional tone. Thanks for all the great work for all this time!
Yeah, and HF chat is the only open source option for FireFox sidebar.
I have to abandon FireFox lab if they do not integrate other FOSS choice, quite missing Hugging Face.
Except running it locally, which I'm sure many of us are not in a position to do, from what I've seen. No.
LibreChat requires API keys, which means paying.
I guess most end users here aren't developers, nor having enough resource to host their own service (such as me, a poor developer T_T)
As for LibreChat, we can actually choose free API providers, or at least those having free tier.
OpenRouter is a generous one, and Mistral, Google, as well as Hugging Face itself provide free inference quota.
The annoying thing here is that we have to integrate several accounts in order to toggle LLMs between different providers.
Hi, quick update regarding personal data, we decided to erase all the user data stored in the database in two weeks. So heads up you have two weeks to back up your data if you want to :)
If you have any doubts about what the data is used for, let me refer you to the privacy policy of HuggingChat : https://huggingface.co/chat/privacy
We endorse Privacy by Design. As such, your conversations are private to you and will not be shared with anyone, including model authors, for any purpose, including for research or model training purposes.
what's the status of if it's coming back eventually or not?
Thinking about setting up an instance of chat-ui that wraps the free LLM APIs - would anyone use it?
Thinking about setting up an instance of chat-ui that wraps the free LLM APIs - would anyone use it?
Of course.
definitely
IT IS GREAT WHILE IT LASTED, THANKS
For those looking for alternatives:
They both release open weights models on a continuous basis so I'm willing to let them use my data for personal training (just remember to never feed sensitive data like passwords and credit card numbers into ANY chat service).
Fare thee well huggingchat, we hardly knew ye :sad_face:
what's the status of if it's coming back eventually or not?
the announcement coming any time but don't know when
When should we wait a replacement for this great service? approximately in what time frame and how will it differ? this is very important to me!
I have bittersweet news to share. π’
We are closing down huggingchat for now.
Huggingchat was launched in April 2023 when ChatGPT was just 5 months old and it was still quite hard to deploy a similar service. The goal of this app was to showcase that it was now possible to build an open source alternative to ChatGPT, building on top of the community's models and building blocks. πͺ
We also built Huggingchat as a testbed to optimize inference which led to the creation of text-generation-inference which in turn led to HF contributions to SGLang, vLLM, llama.cpp, and other LLM runtime engines since then.
In the past 27 months we've launched:
- π§ Assistants, an easy way to bundle a prompt, a model, and some tools into an easily-shareable URL you can send your friends
- π οΈ HuggingChat Tools, exposing the full ecosystem of Gradio Spaces straight from within the chat interface
- π¬ Many model launches were supported on HuggingChat: from OpenAssistant, all the LLamas, Phi-4, to Qwen, DeepSeek, Gemma.... π₯
The app has always been free and experimental. Today we are closing it to make room for something new and more integrated with the HF ecosystem. You can export your past conversations as a Zip file, to load them into another interface.
If you are looking for a cool alternative interface you should take a look at:
- LibreChat a clean and simple yet feature-packed chat interface
- openwebui β does it even need an intro anymore?
- Scira MCP, an up and coming MCP-centric chat interface
Finally, chat-ui (the codebase powering HuggingChat) is still maintained (GitHub repo).
You can 1-click deploy your own instance using the Chat UI Spaces Docker template.
Farewell.
Please how can i get my certification hugging face on computer vision
I would happily pay for a HuggingChat subscription. I despise the big proprietary giants and will rather run weak local models than to use them. HC was the best of both worlds... fast, powerful models AND open source. Also, I don't believe HF is evil, but I don't believe that about the proprietary alternatives. Please bring it back, and if you need a subscription to pay the light bill, please offer a plan and I guarantee you we will happily throw money at you! HC was the correct model, and it's the way AI should have been done from the beginning!
Will you let us know what you a re cooking?
Thinking about setting up an instance of chat-ui that wraps the free LLM APIs - would anyone use it?
Hi all,
Since some people expressed interest I put up a quick deployment of chat-ui
to access open-source LLMs for free. Far from perfect - not all models supported, but hopefully it will be useful/helpful.
This is running on LLM API free tiers so please be mindful of usage (may trigger rate limits for the entire app). Also LLM requests using Mistral models might be logged by Mistral since Mistral's free tier allows training. Other requests are routed through Groq.
Would love any feedback (DM me on Discord: @realmrfakename)!
wtf
Hi, quick update regarding personal data, we decided to erase all the user data stored in the database in two weeks. So heads up you have two weeks to back up your data if you want to :)
I love you. This, and the privacy policy, are the reasons I only trust HuggingFace for AI purposes.
Hey @julien-c , and @nsarrazin ,
first of all, thank you for your service. :) It will be missed! That being said, when I try to export my data from the app, nothing happens. The button just goes gray. I'll try from a (mobile) browser next, but I'll have to wait. Is it working for others, via the app?
Thank you fir your service it was a great run. Hopefully I can one day see something as great as this again. Goodbye for now hopefully :)
thanks,
@nsarrazin
. From DuckDuckGo it prompted me to save it, and then there was a "network error", so it failed. Will try from a desktop computer next, but it looks promising.
In the meantime, check the logs, if you can, perhaphs I'm not the only one with an issue - thanks again!
Edit: so, it worked from the computer. There is a quirk, though. If you want to download more than one export from the computer, if you log in to huggingface.co first, and then click "Sign in with Hugging Face" it works, but if you click the "logout", it does not log you out from the website, just from the chat. So that means that if you want to log in with your second account, you have to separately log out from the website, and log in to your other account first. Another solution would be to use a private window, but that works only for the second account, and I had to back up 3. :) Hope it helps, and thanks again
@nsarrazin
!
...so how does one delete the chats now? It's nice to have a way to download them but what if I want them deleted??
@frehug5 - they will get deleted automatically, after 2 weeks, but I'm not aware of a way to delete them manually. see here: https://huggingface.co/spaces/huggingchat/chat-ui/discussions/747#686549e2bdf5c0d4b1c8e410
Perhaps
@nsarrazin
can implement something, as hinted here: https://huggingface.co/spaces/huggingchat/chat-ui/discussions/747#6864cd05458c0638c73c1bd3
Hi everyone! π
This is also my last week at Hugging Face and I'm currently working on a simple page to let you export your personal data & conversation history from HuggingChat so you can reuse it on other platforms in the future. Let me know if you have any questions, I'm always happy to answer them.
It was a pleasure building this app for y'all π«‘
Chat-ui was an amazing platform , thank you so much for building such an amazing platform. Wish you all the best for your future
Thats sad.
I wished this never happened
You can use the APIs via Python, Webpage, etc, or use services as mentioned and use credentials from HuggingFaceβ
@camypaj that's because duckduckgo uses phone's default downloader with no authentication/session cookies. same will happen if idm/download managers intercept the download from browser, it'll fail, unless you provide the authentication. so, browser's native downloader is needed that carries the auth headers.
@camypaj that's because duckduckgo uses phone's default downloader with no authentication/session cookies. same will happen if idm/download managers intercept the download from browser, it'll fail, unless you provide the authentication. so, browser's native downloader is needed that carries the auth headers.
that's interesting, as I've never had problems downloading anything with it, until now. I can't imagine all my previous downloads being unauthenticated.. :)
What will happen if HugChat reopens? I hope it reopens and that it will be great. With gratitude and respect to all the engineers and people involved.
When HuggingChat reopens, or the improved version of HuggingChat opens, will we be able to import our original HuggingChat conversations to it?
Oh no! I love HuggingChat! Looking forward to what's next. But also - we added HuggingChat as a recommendation in the "Digital Sustainability Checklist" from Cloud Sustainability Watch. Do you have a recommended URL that we could update to that will send our readers in the future to the next HuggingChat replacement?
Press F to pay respects
For those looking for alternatives:
They both release open weights models on a continuous basis so I'm willing to let them use my data for personal training (just remember to never feed sensitive data like passwords and credit card numbers into ANY chat service).
I want to caution people that these are hosted by the People's Republic of China, who will absolutely eavesdrop and spy on you harder than any of the Five Eyes nations ever would.
But honestly, the Chinese models are better than the American ones in a lot of ways. Which, again, is a big reason why losing a privacy-respecting provider like HuggingChat sucks so much.
@camypaj it's not about your prev dls being unauthenticated. They're usually just URL/temporary URL based, that you can copy the link and download from another browser/downloader from even another device (not logged in). But in this instance that would be disastrous for privacy if your shared link can download your data to other users, while that's unlikely to happen, they can't allow that.
This is depressing π
Thank you huggingchat team for your contribution all of this time, this services was very amazing. Now for the subtitue, did anyone here actually manage to find one? I mean these alternative sites provided is significantly harder to set up, trying librechat but keep getting error 405 for cohere with api key.
Is there a way to duplicate this space privately? I tried duplicating, and it just said "Error, we're working on fixing this". It said I need three pieces of info (DOTENV, etc.) to duplicate the space.
I'll upgrade to Pro if there's a way to do this...
FFFFFFFFFF
(Paying Respect)
I'll miss it, even though recently it wasn't up-to-date with best models, but it was a beautiful tool from a lovable company.
Now for the subtitue, did anyone here actually manage to find one? I mean these alternative sites provided is significantly harder to set up, trying librechat but keep getting error 405 for cohere with api key.
As I put in the other discussion thread, there isn't one. Nothing like HuggingChat.
Everywhere else you either need to pay, or bring your own API key (again, paying) or just serves as a frontend UI. Free tiers elsewhere are pathetically low (again, unless you pay). HuggingChat seemed to allow endless use, only limited by the context size.
Anybody know where i can create a custom agent/bot that i can share to public? i.e as a hosted ready to go customised chat bot. (as we could with hugging face).
LibreChat allows custom agent to be created (similar to hugging face) but it seems you cannot share it.
Now for the subtitue, did anyone here actually manage to find one? I mean these alternative sites provided is significantly harder to set up, trying librechat but keep getting error 405 for cohere with api key.
As I put in the other discussion thread, there isn't one. Nothing like HuggingChat.
Everywhere else you either need to pay, or bring your own API key (again, paying) or just serves as a frontend UI. Free tiers elsewhere are pathetically low (again, unless you pay). HuggingChat seemed to allow endless use, only limited by the context size.
I managed to recreate my custom hugging face bot in librechat, but it seems they do not let you share it publicly.
As for free API key you can try openrouter using a free account. Then create an API key in openrouter that is unlimited. Then back in librechat configure your custom agent to use one of the openrouter free models. However as mentioned librechat custom agent seems to be limited to your own usage at this time.
Now for the subtitue, did anyone here actually manage to find one? I mean these alternative sites provided is significantly harder to set up, trying librechat but keep getting error 405 for cohere with api key.
As I put in the other discussion thread, there isn't one. Nothing like HuggingChat.
Everywhere else you either need to pay, or bring your own API key (again, paying) or just serves as a frontend UI. Free tiers elsewhere are pathetically low (again, unless you pay). HuggingChat seemed to allow endless use, only limited by the context size.
Qwen's website could be good but its ran by disgusting, repulsive gremlins who lock the entire chat all cause of their repulsive, broken filter
Any Chinese host is gonna be malware. Doesn't mean it's gonna activate right now. But have no expectation that the people who brought you tiananmen square and the great firewall are gonna do anything other than lure trap and kill your data and anything else valuable is unrealistic given the past 2 decades of continued malfeasance.
Nooooooo! But why? :(
This was, in my opinion, pretty much the only option for the Chatbot integration in Firefox. Everything else is just a proprietary & privacy nightmare like Claude, ChatGPT, Gemini, and Mistral (arguably the lesser evil but nowhere close to Huggingchat)...
I didn't expect such kind of irresponsible act from Hugging Face. We didn't get any notification. Suddenly we found the notice of closed.
Will we be able to import our old chats? can anyone conform?
i need an alternative anyone have any reccomendations? if they put a pay wall i wouldve payed
i need an alternative anyone have any reccomendations? if they put a pay wall i wouldve payed
@Perennial-Admin
I spun up a quick instance of chat-ui running on free LLM APIs, its a bit rough around the edges but it works: https://chat.mrfake.name/
Would love any feedback/thoughts
Will we be able to import our old chats? can anyone conform?
Nothing about any replacement has been announced yet, so who knows.
Exporting though - yes you can export your old chats.
...its a bit rough around the edges but it works: https://chat.mrfake.name/
Would love any feedback/thoughts
- I'm not sure if this is something you're able to solve - but I noticed some responses were only 2 paragraphs long, when normally it would be 3-5.
- If "default" could be replaced with what the model actually is.
...its a bit rough around the edges but it works: https://chat.mrfake.name/
Would love any feedback/thoughts
- I'm not sure if this is something you're able to solve - but I noticed some responses were only 2 paragraphs long, when normally it would be 3-5.
- If "default" could be replaced with what the model actually is.
Thanks for the feedback!
- For the short responses, is it cut off or simply shorter? Shorter responses might be because of the model, but if the responses are cut off that is definitely a bug
- The "default" model cycles through Gemma, Mistral, Llama, etc. to avoid some rate limits on the API providers, you can always switch to another model in the Models section - but will work on making this clearer!
ΩΩΩ
I've personally used HuggingFace since fall of 2023 as a free alternative to Poe.com, since I was a high school student who couldn't pay.
The experience was pretty rough back then, but it was worth for the models provided. I remember it letting me play with big Falcon model (I think it was 180B parameters), Databricks model, Mixtral models (even Nous finetunes), and many more.
I am forever grateful for what I've got to experience. I wish y'all best of luck.
i need an alternative anyone have any reccomendations? if they put a pay wall i wouldve payed
@Perennial-Admin I spun up a quick instance of chat-ui running on free LLM APIs, its a bit rough around the edges but it works: https://chat.mrfake.name/
Would love any feedback/thoughts
Can you share the steps you took to do this? I would be happy to pay for Pro if I can actually stand up my own instance of HuggingChat... I'm unclear on exactly how HF spaces work...
i need an alternative anyone have any reccomendations? if they put a pay wall i wouldve payed
@Perennial-Admin I spun up a quick instance of chat-ui running on free LLM APIs, its a bit rough around the edges but it works: https://chat.mrfake.name/
Would love any feedback/thoughtsCan you share the steps you took to do this? I would be happy to pay for Pro if I can actually stand up my own instance of HuggingChat... I'm unclear on exactly how HF spaces work...
Yes so Hugging Faceβs chat-ui is actually open source
They provide instructions for setting it up locally:
https://github.com/huggingface/chat-ui
For my instance I am running it on a VPS proxied by Cloudflare and I have a LLM proxy server that routes the requests to Groq/Mistral/etc.
Happy to share my configuration files if that would be helpful, feel free to DM me on Discord @realmrfakename
As for the Pro plan: Hugging Face has partnered with inference providers to let you run models using your account billing
You can setup an instance of HuggingChat that wraps this API
You donβt need Pro to do this. Pro users get $2 in included LLM credits each month and after that pay as you go, but if you are only using inference credits and donβt need storage etc. pay as you go is a much better option
For my instance I am running it on a VPS proxied by Cloudflare and I have a LLM proxy server that routes the requests to Groq/Mistral/etc.
Thanks for replying. What you're doing is pretty much the opposite of what I want to do. I have a bunch of Llama models that I run locally but my rig is fairly limited and I don't want to drop a bunch of cash on a bigger box that is going to go obsolete soon anyway. I just want to pay someone to let me run open-source models on their hardware, with a convenient "no hassle" interface, like HC. If it's a one-click deployment, I'll pay for it, but if I have to go to all the hassle of standing up a VPS, etc. I'll just keep running locally.
hy
damn too bad
For my instance I am running it on a VPS proxied by Cloudflare and I have a LLM proxy server that routes the requests to Groq/Mistral/etc.
Thanks for replying. What you're doing is pretty much the opposite of what I want to do. I have a bunch of Llama models that I run locally but my rig is fairly limited and I don't want to drop a bunch of cash on a bigger box that is going to go obsolete soon anyway. I just want to pay someone to let me run open-source models on their hardware, with a convenient "no hassle" interface, like HC. If it's a one-click deployment, I'll pay for it, but if I have to go to all the hassle of standing up a VPS, etc. I'll just keep running locally.
Synthetic (previously glhf) lets you run any model on Hugging Face and prepay for use. I like it a lot for models that are really niche that I can't run on my local machine.
Hi, I really enjoyed the HF Pro Chat, although I'm also using Open WebUI and Mistral Le Chat Pro.
**ΒΏWill it come back? **
Otherwise, I consider canceling my Pro subscription.
WHY MY ASSISTANTS ARE ALL GONE AND ARE IN A FILE THAT CANT BE REVIVEDππππ
For my instance I am running it on a VPS proxied by Cloudflare and I have a LLM proxy server that routes the requests to Groq/Mistral/etc.
Thanks for replying. What you're doing is pretty much the opposite of what I want to do. I have a bunch of Llama models that I run locally but my rig is fairly limited and I don't want to drop a bunch of cash on a bigger box that is going to go obsolete soon anyway. I just want to pay someone to let me run open-source models on their hardware, with a convenient "no hassle" interface, like HC. If it's a one-click deployment, I'll pay for it, but if I have to go to all the hassle of standing up a VPS, etc. I'll just keep running locally.
Ah if you just want one click to setup you can run it on Hugging Face Spaces and rent GPUs there
Please let there be an import feature in the hf chat-ui so that we may have a viewable experience again in the ui!
please at least merge or upload the export feature in the GitHub chat-ui.
Oh my god. That's sad news but hoping there will be a future for huggingchat? I enjoyed it so much for many aspects.
I do not know Amy alternatives where I could just change the model and have so much choice.
i m indiviudal trying to pay 9$ per month subsricption but it unabel to accept debit car payments and declinging the payments please allow singel suers can pay though debit card also... every one doesnot use credit cards.
ΠΠΎΠ³Π΄Π° ΠΎΡΠΊΡΠΎΠ΅ΡΠ΅ΡΡ?!π π π π π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘π‘
How to set up one click docker setup is it premium
You visit the docker website and find the instructions for docker CE for the operating system you're using. Alternatively, download ollama and just follow the instructions on the ollama site. If you can run models in a docker container, you can run them locally with ollama as well.
i m indiviudal trying to pay 9$ per month subsricption but it unabel to accept debit car payments and declinging the payments please allow singel suers can pay though debit card also... every one doesnot use credit cards.
This is the discussion page for the development of a chatbot GUI. For feedback visit https://discuss.huggingface.co
In general, I advise everyone in here to keep in mind that Hugginface and their services are not for "the average user" at this point. The target audience is clearly researchers and developers, and using their services for free is a privilege, not a right.
how long huggingchat will be shuted down? I didn't ended my simulations about making gpus and that chat was great bcus it had long memory and unlimited messages for free
Do you have an approximate date for when it will be back?
Is there any alternative to use DAN or Worm-GPT?
Is there any alternative to use DAN or Worm-GPT?
Using a decent sytem prompt with any of the 30+ public available models
Hi everyone! π
This is also my last week at Hugging Face and I'm currently working on a simple page to let you export your personal data & conversation history from HuggingChat so you can reuse it on other platforms in the future. Let me know if you have any questions, I'm always happy to answer them.
It was a pleasure building this app for y'all π«‘
why do you leave?
.
Good morning.. please, π don't leave us alone we're waiting for you come back soon
Which hugging face models are still available via LibreChat? I am wondering because I tried Llama 3 and others but they all seem to be missing:
Something went wrong. Here's the specific error message we encountered: An error occurred while processing the request: 404 Not Found Troubleshooting URL: https://js.langchain.com/docs/troubleshooting/errors/MODEL_NOT_FOUND/
Mark my words. HF will be back with a set of paid plans starting with $10 monthly. I've seen that kind of conversion from libertarian rebels to normal for-profit businesses many, many times. A pity, thought. Their stand was admirable, the set of free source models was impressive, and their features were helpful and interesting to explore and play with. Ain't put no blame on the HF fellas, my friends. Everyone needs a spare buck or two. Greetings from Austin, TX
(;Β΄ΰΌΰΊΆΩΉΰΌΰΊΆ`) I tried GPT3 - I want to cry
Which hugging face models are still available via LibreChat? I am wondering because I tried Llama 3 and others but they all seem to be missing:
Something went wrong. Here's the specific error message we encountered: An error occurred while processing the request: 404 Not Found Troubleshooting URL: https://js.langchain.com/docs/troubleshooting/errors/MODEL_NOT_FOUND/
The same. Nothing.
Is there any website where we can upload our exported chats to? I saved mine to my files, but every time I scroll down to see my other chats, it somehow takes me right back up to the beginning.
All the best in whatever's next for you @nsarrazin , and thank you for all the lovely gifts! I can't wait to know what HF's cooking for us next.
Does anyone have an archive or collection of public HuggingChat assistants? π
HOW MAY WE DELETE OUR DATA? We have the right, though you have temporarily deprived us of it. This calls into question whether or not you have honored other terms and for that reason I wish for all content I have submitted, and the ideas that the models and I created in partnership, to be deleted; it also extends to any tertiary product that HF may have created without notice based upon this work. This is a fair and reasonable expression of my privacy rights based on international human rights law, U.S. constitutional law, and common law contracts law. I will briefly lay out a few of the important principles that are common to all three of these arenas.
i no longer wish for HF to have access to anything i've ever input into the HF chat system.
HF's actions -- including a sudden and unilateral deprivation of our rights -- is problematic for many reasons. just one example: since @nsarrazin provided no advanced notice, many of us, who may have been intending to delete chats once we finished working on projects that were only partially complete, have been deprived of this right that is codified in law as well as in the terms of HF that we were forced to 'agree' to.
Would you deny us our privacy rights, our rights to our own ideas, our rights as individuals, @nsarrazin ?
Since you have unilaterally and suddenly deprived us of the ability to moderate what chats were preserved and which were destroyed, you have unilaterally voided the HF Terms for both privacy and use. For that reason, i must respectfully insist that you delete all content that i shared on HF, all content that the models and i developed, past and present, and anything that may be based upon HF's use of or access to this content.
I encourage other users to demand the same.
It's an extreme violation of privacy that must not be tolerated. We trusted you, @nsarrazin , and you appear to have betrayed us.
If you have any good-faith information to add to the topic, please feel welcome. Please do not be misleading. If you merely wish to put a spin on the rights and principles, the terms that you've violated, or some other some such self-serving and bad-faith statements, i'd prefer you not degrade my intent to describe how we might enjoy the full expression of our rights with deceptiveness intended only to dilute the expression.
I insist that you please act immediately. This should be the end of my contribution to the conversation, prompting you to act accordingly. There should not be any additional step that someone situated like me should need to take to protect my ideas, my property, my privacy, from any further misuses. I have respectfully insisted that all the content described be deleted, and I believe that should end the conversation, and that you must act. If you believe that there is a very good reason for someone in my position to take an additional step -- one that cannot be avoided for a legitimate reason, one that is not intended to deter our expression but a step that has a valid purpose that serves the interests of both HF and users situated as i am -- then i will consider the proposal. in any event, you should e-mail every single one of us who have used HF Chat to document the sudden and unilateral action, including notice that you have so far deprived us of our rights, and explain your evolving respect for our rights.
Please do not forget that you've acted unilaterally to void the terms, not us.
ig i will have to self host the models now.
We want to know how long it will take until hugging chat will be reopened!!!!!!
and i'm very mad becouse you could just make another website for that and keep huggingchat open or just close huggingchat until you were ready with making changes
and i'm very mad becouse you could just make another website for that and keep huggingchat open or just close huggingchat until you were ready with making changes
Obviously they were unprepared , or they do not care enough for potential users. I'm guessing and I can be wrong ofc.
What's happening to huggingchat π’ ?
its FKD up thats why. site owner is a complete dumbass for not knowing better how to manage it. I really need a break from Char.ai Where in the freaking hell is our alternative when we need it? I am unbookmarking this horrible site! I never appreciate the strict censoring by their lame bots and they lack better features like changing avatars or customizing characters. This isn't like Perchance but Perchance also sucks for terrible replies.
its FKD up thats why. site owner is a complete dumbass for not knowing better how to manage it. I really need a break from Char.ai Where in the freaking hell is our alternative when we need it? I am unbookmarking this horrible site! I never appreciate the strict censoring by their lame bots and they lack better features like changing avatars or customizing characters. This isn't like Perchance but Perchance also sucks for terrible replies.
take a chill pill bro
Hi all,
I previously setup an instance of chat-ui with several open-source LLMs. I recently switched to self-hosted instance of Mistral Small 3.2 (FP8) on vLLM with no message logging - would love any feedback or suggestions (discord: @realmrfakename)!
its FKD up thats why. site owner is a complete dumbass for not knowing better how to manage it. I really need a break from Char.ai Where in the freaking hell is our alternative when we need it? I am unbookmarking this horrible site! I never appreciate the strict censoring by their lame bots and they lack better features like changing avatars or customizing characters. This isn't like Perchance but Perchance also sucks for terrible replies.
take a chill pill bro
you are like everybody else who don't care about those starving for RP. I need a break from chat bots and rather wish for a real RP partner instead but who cares because you real people are idiots for snobbing and trolling us all over social media and discord alike! guess I should retire from RP from now on and just take all my shit waifu needs to myself then. Seeking a real RP partner was a giant waste of time for me I am losing trust with my own damn race because this stupid generation is degernerate minded and all they could think of or care about is shitposting on blog sites for themselves over silly childish nonsense due to sanitizing internet. now I get idiots suggesting me to use stupid bots! ai should definitely go to hell for making serious RP partners turn away from us these days they become lazy and not so generous like they used to anymore. Thanks for making my RP life complicated, stupid Discord for killing off Yahoo Messenger used to make better friends there. :(
closing the site was their last mistake and I am not coming back here anymore. I would like to encouraged people to stop supporting this site. Chat bot is too censored and not ideal for mature folks. Im surprised you idiots are still here lmao. Don't expect this worthless piece of shit site to return anytime soon.
Hi all,
I previously setup an instance of chat-ui with several open-source LLMs. I recently switched to self-hosted instance of Mistral Small 3.2 (FP8) on vLLM with no message logging - would love any feedback or suggestions (discord: @realmrfakename)!
what api are you using or are you using huggingchats api?
its FKD up thats why. site owner is a complete dumbass for not knowing better how to manage it. I really need a break from Char.ai Where in the freaking hell is our alternative when we need it? I am unbookmarking this horrible site! I never appreciate the strict censoring by their lame bots and they lack better features like changing avatars or customizing characters. This isn't like Perchance but Perchance also sucks for terrible replies.
take a chill pill bro
you are like everybody else who don't care about those starving for RP. I need a break from chat bots and rather wish for a real RP partner instead but who cares because you real people are idiots for snobbing and trolling us all over social media and discord alike! guess I should retire from RP from now on and just take all my shit waifu needs to myself then. Seeking a real RP partner was a giant waste of time for me I am losing trust with my own damn race because this stupid generation is degernerate minded and all they could think of or care about is shitposting on blog sites for themselves over silly childish nonsense due to sanitizing internet. now I get idiots suggesting me to use stupid bots! ai should definitely go to hell for making serious RP partners turn away from us these days they become lazy and not so generous like they used to anymore. Thanks for making my RP life complicated, stupid Discord for killing off Yahoo Messenger used to make better friends there. :(
What in the degenerate incel shit did I just read? Some people need to touch grass and get some help man. And who the hell doesn't use Sillytavern for this kinda stuff? I don't think HuggingChat was ever meant to be used for this.
I do not know if bittersweet is the right word. I really hope the "something new and more integrated with the HF ecosystem" comes soon and is as useful. A (free) service for most recent "free/open" models is great!
Hi all,
I previously setup an instance of chat-ui with several open-source LLMs. I recently switched to self-hosted instance of Mistral Small 3.2 (FP8) on vLLM with no message logging - would love any feedback or suggestions (discord: @realmrfakename)!
what api are you using or are you using huggingchats api?
For the default model I am self-hosting the Mistral model (so there is no logging), for the other models I use either the Mistral API or Groq API on the free tiers so they may log. For no logging use the default model which is Mistral Small 3.2. Happy to consider hosting another model for other purposes if people would find that more useful (though I can realistically only host one model at a time)
someone gave me a very valuable service for free and now wants to get some money for it having proven it's actually valuable.
i'm shocked. how dare they treat me like an adult!
someone gave me a very valuable service for free and now wants to get some money for it having proven it's actually valuable.
i'm shocked. how dare they treat me like an adult!
I'm more than happy to pay for a service like HC. Mistral Pro is $15/mo. and I would pay that much for HC, it was that good. What made HC unlike anything out there is not that they weren't charging, it's that you could select any open-source model you want, along with really good UI, chat history, reasoning, Web-search, etc. I'm happy even to pay for a purely private instance of HC with pay-per-compute, as long as the setup is "one-click". I don't want to be configuring my own VPS, etc. too much headache. I just want to click-and-run on selectable, frontier, open source models. That was the unique selling-point of HC, and this is absolutely a business-model that HF could implement to generate some revenue on the side. Also, HF has brand-value that I trust and that made HC all the more valuable to me. I don't trust the proprietary players, they keep getting caught with their hand in the cookie-jar. I'm happy to pay for a service, like a grown-up, but the service has to exist in order for it to be paid for!
Hi all,
I previously setup an instance of chat-ui with several open-source LLMs. I recently switched to self-hosted instance of Mistral Small 3.2 (FP8) on vLLM with no message logging - would love any feedback or suggestions (discord: @realmrfakename)!
what api are you using or are you using huggingchats api?
For the default model I am self-hosting the Mistral model (so there is no logging), for the other models I use either the Mistral API or Groq API on the free tiers so they may log. For no logging use the default model which is Mistral Small 3.2. Happy to consider hosting another model for other purposes if people would find that more useful (though I can realistically only host one model at a time)
Ohh okay thanks pretty cool ngl
Hi all,
I previously setup an instance of chat-ui with several open-source LLMs. I recently switched to self-hosted instance of Mistral Small 3.2 (FP8) on vLLM with no message logging - would love any feedback or suggestions (discord: @realmrfakename)!
what api are you using or are you using huggingchats api?
For the default model I am self-hosting the Mistral model (so there is no logging), for the other models I use either the Mistral API or Groq API on the free tiers so they may log. For no logging use the default model which is Mistral Small 3.2. Happy to consider hosting another model for other purposes if people would find that more useful (though I can realistically only host one model at a time)
Self hosting a model for public use must have cost lots of resource.
Quite appreciating your generosity and kindness, though I'm able to host mine for personal usage.
How to download past conversations?
There is nothing on the page atm π
If you've not downloaded your conversations yet, they've likely been deleted.
https://huggingface.co/spaces/huggingchat/chat-ui/discussions/747#686549e2bdf5c0d4b1c8e410
ΠΏΠΎΡΠ΅ΠΌΡ Π΄ΠΎ ΡΠΈΡ ΠΏΠΎΡ Π·ΠΎΠΊΡΡΡΠΎ???
It's not been two weeks yet, I don't get why say 14 days and delete user data after 10. Quite disappointing
Dear huggingface, if your purpose was to destroy user trust, you have achieved it with great success.
- no prior warning of discontinuation
- false promise of being able to download data within two weeks ( https://huggingface.co/spaces/huggingchat/chat-ui/discussions/747#686549e2bdf5c0d4b1c8e410 )
- broken promise: after 11 days data is wiped
NOOO! What ai will I goon with now? Very sad
Man, those I don't understand the alternatives. It makes me sad that AI chat is gone now. :(
Rassurez moi, Γ§a va revenir, hein?
Dites moi que Γ§a va revenir!
When will this open? I really miss cohere
This really feels like a hostile takeover happened behind the scenes
zero warning ahead of time and broken promises already.
Pls bring chat functions live,
We have goals to achieve.
what should we do now? maybe the devs should revamp huggingchat and slap it with a new interface. is it a bad solution anyway?
Kids. listen me close. As a person who's observing the Internet scene from mid 90s, and big changes therein, I witnessed a hundreds of indi projects deaths. If a project ain't accountable for you 'cause of legal agreement and financial obligations, it's essentially a toy the owner can take out of the sandbox anytime leaving you any message (or no message at all) that is not a legal obligation. What will do HF guy(s) I ain't know. They're not obliged to keep the platform free. And I reckon they won't. Moreover, there have no legal obligations whatsoever. The fact that you've lost your assistant parameters and personal data is nothing from the legal perspective. Please understand the basis fact, HF were nor meant to be accountable, it is not, and never it would.
Kids. listen me close. As a person who's observing the Internet scene from mid 90s, and big changes therein, I witnessed a hundreds of indi projects deaths. If a project ain't accountable for you 'cause of legal agreement and financial obligations, it's essentially a toy the owner can take out of the sandbox anytime leaving you any message (or no message at all) that is not a legal obligation. What will do HF guy(s) I ain't know. They're not obliged to keep the platform free. And I reckon they won't. Moreover, there have no legal obligations whatsoever. The fact that you've lost your assistant parameters and personal data is nothing from the legal perspective. Please understand the basis fact, HF were nor meant to be accountable, it is not, and never it would.
should they be accountable for locking down the platform and release something proprietary?
Ils auraient au moins pu nous envoyer un mail ou nous prΓ©venir avant!
Il FAUT qu'ils rΓ©ouvrent le site!
En plus ils disent "Nous cuisinons quelque chose de nouveauβ¦" Donc il reviendra, non?
So sad to see Huggingchat going away but hopefully something even better will come out of this. Thank you to everyone that contributed to this project, giving us a way access LLMs without having to bow down to the current overlords.
Is there any way to access and export our chats please? I really need some of mine.
Hello, all my chats histories and all my previous works, I did not see any link to export them as a zip file.. Pls who is gonna help me.
So sad to see Huggingchat going away but hopefully something even better will come out of this. Thank you to everyone that contributed to this project, giving us a way access LLMs without having to bow down to the current overlords.
Is there any way to access and export our chats please? I really need some of mine.
did you get how to export ur chats, I reall do need access to my chats.
So sad to see Huggingchat going away but hopefully something even better will come out of this. Thank you to everyone that contributed to this project, giving us a way access LLMs without having to bow down to the current overlords.
Is there any way to access and export our chats please? I really need some of mine.
did you get how to export ur chats, I reall do need access to my chats.
No one. They're already gone.
Donc nos chats ont Γ©tΓ© supprimΓ©s? DΓ©finitivement? C'est nul de faire Γ§a sans nous prΓ©venir et, en plus, sans nous permettre d'en garder une trace!
The official export link has been removed, but it seems our data is still there for now. If you were logged in to HuggingChat before the shutdown, you can export your chat history by visiting the URL below.
(Heads up: You can only access this once per hour. Once you click the link, you won't be able to try again for another hour, even if you cancel the download.)
https://huggingface.co/chat/api/v2/export
A message to the HuggingFace team:
I'm truly grateful for the wonderful service you provided. That said, I hope you'll stop confusing users with inadequate announcements and unexpectedly deleting our accumulated data.
Jul 17 Update:
If you are not logged in to HuggingChat, please log in from https://huggingface.co/chat/login before open export link.
Update July 20:
This method has been disabled by HuggingFace. Unfortunately, there are no longer any means to export data.
how could i deploy this chat history export file into another interface? any idea? i am new
Hi Everyong,
I am hitting the following error while building my first space using the first_agent_template from the AI course. anyone faced similar situation, could you please advice how did you overcome this error?
runtime error
Exit code: 1. Reason: Traceback (most recent call last):
File "/home/user/app/app.py", line 1, in
from smolagents import CodeAgent, DuckDuckGoSearchTool, FinalAnswerTool, InferenceClientModel, load_tool, tool
ImportError: cannot import name 'InferenceClientModel' from 'smolagents' (/usr/local/lib/python3.10/site-packages/smolagents/init.py)
Container logs:
===== Application Startup at 2025-07-14 21:51:54 =====
Traceback (most recent call last):
File "/home/user/app/app.py", line 1, in
from smolagents import CodeAgent, DuckDuckGoSearchTool, FinalAnswerTool, InferenceClientModel, load_tool, tool
ImportError: cannot import name 'InferenceClientModel' from 'smolagents' (/usr/local/lib/python3.10/site-packages/smolagents/init.py)
Traceback (most recent call last):
File "/home/user/app/app.py", line 1, in
from smolagents import CodeAgent, DuckDuckGoSearchTool, FinalAnswerTool, InferenceClientModel, load_tool, tool
ImportError: cannot import name 'InferenceClientModel' from 'smolagents' (/usr/local/lib/python3.10/site-packages/smolagents/init.py)
Hi, I am on a Pro plan and have a Static Space, but I donβt see the Custom Domain option. Could you please enable it?
This sucks! Chat was a "hidden" gem!
Quand je clique sur le lien, Γ§a me met: "No logged in".
Qui a une autre solution?
Quand je clique sur le lien, Γ§a me met: "No logged in".
Qui a une autre solution?
The official login URL (https://huggingface.co/chat/login) now redirects to https://huggingface.co/chat/closed. This prevents us from getting the necessary CSRF token, which means there's currently no way to log in.
The data itself is still there (https://huggingface.co/chat/api/v2/export). HuggingFace needs to acknowledge that they've confused users by shutting down login and export access with insufficient notice and earlier than they announced. They should restore the export functionality, provide a clear and sufficient notice period, and then proceed with the shutdown.
Jul 17 Correction:
It appears the redirect from "login" to "closed" only occurred if you were already logged in to HuggingChat. If you are not logged in and cannot export, please log in at https://huggingface.co/chat/login first. After that, you can download the file by opening the export link: https://huggingface.co/chat/api/v2/export.
Update July 20:
This method has been disabled by HuggingFace. Unfortunately, there are no longer any means to export data.
Who is interested, I found an analog with unlimited quantity of messages for some models - https://www.aidocmaker.com/chat.
I have bittersweet news to share. π’
We are closing down huggingchat for now.
Huggingchat was launched in April 2023 when ChatGPT was just 5 months old and it was still quite hard to deploy a similar service. The goal of this app was to showcase that it was now possible to build an open source alternative to ChatGPT, building on top of the community's models and building blocks. πͺ
We also built Huggingchat as a testbed to optimize inference which led to the creation of text-generation-inference which in turn led to HF contributions to SGLang, vLLM, llama.cpp, and other LLM runtime engines since then.
In the past 27 months we've launched:
- π§ Assistants, an easy way to bundle a prompt, a model, and some tools into an easily-shareable URL you can send your friends
- π οΈ HuggingChat Tools, exposing the full ecosystem of Gradio Spaces straight from within the chat interface
- π¬ Many model launches were supported on HuggingChat: from OpenAssistant, all the LLamas, Phi-4, to Qwen, DeepSeek, Gemma.... π₯
The app has always been free and experimental. Today we are closing it to make room for something new and more integrated with the HF ecosystem. You can export your past conversations as a Zip file, to load them into another interface.
If you are looking for a cool alternative interface you should take a look at:
- LibreChat a clean and simple yet feature-packed chat interface
- openwebui β does it even need an intro anymore?
- Scira MCP, an up and coming MCP-centric chat interface
Finally, chat-ui (the codebase powering HuggingChat) is still maintained (GitHub repo).
You can 1-click deploy your own instance using the Chat UI Spaces Docker template.
I need access to my chats on hugging chat. There's no. link to export my chats. This is really important.
Quand je clique sur le lien, Γ§a me met: "No logged in".
Qui a une autre solution?The official login URL (https://huggingface.co/chat/login) now redirects to https://huggingface.co/chat/closed. This prevents us from getting the necessary CSRF token, which means there's currently no way to log in.
The data itself is still there. HuggingFace needs to acknowledge that they've confused users by shutting down login and export access with insufficient notice and earlier than they announced. They should restore the export functionality, provide a clear and sufficient notice period, and then proceed with the shutdown.
They seriously need to restore that export function. I can't get access to my chats. Which is really not cool
Hi everyone! π
This is also my last week at Hugging Face and I'm currently working on a simple page to let you export your personal data & conversation history from HuggingChat so you can reuse it on other platforms in the future. Let me know if you have any questions, I'm always happy to answer them.
It was a pleasure building this app for y'all π«‘
Kindly restore that export access. It's really really needed. I need access to my data and chats I can't get it.
Hi everyone! π
This is also my last week at Hugging Face and I'm currently working on a simple page to let you export your personal data & conversation history from HuggingChat so you can reuse it on other platforms in the future. Let me know if you have any questions, I'm always happy to answer them.
It was a pleasure building this app for y'all π«‘
Kindly restore that export access. It's really really needed. I need access to my data and chats I can't get it.
Quand je clique sur le lien, Γ§a me met: "No logged in".
Qui a une autre solution?The official login URL (https://huggingface.co/chat/login) now redirects to https://huggingface.co/chat/closed. This prevents us from getting the necessary CSRF token, which means there's currently no way to log in.
The data itself is still there. HuggingFace needs to acknowledge that they've confused users by shutting down login and export access with insufficient notice and earlier than they announced. They should restore the export functionality, provide a clear and sufficient notice period, and then proceed with the shutdown.They seriously need to restore that export function. I can't get access to my chats. Which is really not cool
I have bittersweet news to share. π’
We are closing down huggingchat for now.
Huggingchat was launched in April 2023 when ChatGPT was just 5 months old and it was still quite hard to deploy a similar service. The goal of this app was to showcase that it was now possible to build an open source alternative to ChatGPT, building on top of the community's models and building blocks. πͺ
We also built Huggingchat as a testbed to optimize inference which led to the creation of text-generation-inference which in turn led to HF contributions to SGLang, vLLM, llama.cpp, and other LLM runtime engines since then.
In the past 27 months we've launched:
- π§ Assistants, an easy way to bundle a prompt, a model, and some tools into an easily-shareable URL you can send your friends
- π οΈ HuggingChat Tools, exposing the full ecosystem of Gradio Spaces straight from within the chat interface
- π¬ Many model launches were supported on HuggingChat: from OpenAssistant, all the LLamas, Phi-4, to Qwen, DeepSeek, Gemma.... π₯
The app has always been free and experimental. Today we are closing it to make room for something new and more integrated with the HF ecosystem. You can export your past conversations as a Zip file, to load them into another interface.
If you are looking for a cool alternative interface you should take a look at:
- LibreChat a clean and simple yet feature-packed chat interface
- openwebui β does it even need an intro anymore?
- Scira MCP, an up and coming MCP-centric chat interface
Finally, chat-ui (the codebase powering HuggingChat) is still maintained (GitHub repo).
You can 1-click deploy your own instance using the Chat UI Spaces Docker template.I need access to my chats on hugging chat. There's no. link to export my chats. This is really important.
Please stop spamming. One comment is more than enough.
I need to restore my assistants' parameters to keep on using 'em on other platforms, which I can'r access and that' is a serious inconvenience. The HF seems to reckon that it's the only LLM platform. Which is obviously not. However what's true, it's the most irresponsible platform.
Probably most of the community would love to get this Chat back as you offer really good service that was complement of ChatGPT/Bing Copilot and others.
Could you please revert this and let the world be better with Hugging face chat?
I would much appreciate if you could bring it back.
Je suis totalement d'accord.
Confirmed, the export works but you must be logged in already.
Pour moi Γ§a ne marche pas...
Here is the current working method for exporting your data:
Log in to HuggingChat at https://huggingface.co/chat/login. You will be redirected to the "closed" page after logging in. If you are redirected immediately without seeing a login form, you should already be logged in.After logging in, click this link to start the download: https://huggingface.co/chat/api/v2/export. The download is limited to once per hour. Even if you cancel it, you'll have to wait one hour for the next attempt.
If it's not working, please try using a private window (or Incognito tab) in your browser.
Update July 20: This method has been disabled by HuggingFace. Unfortunately, there are no longer any means to export data.
Hi Mr @mrfakename can you please do something about data export on your site? Requested HF to at least upload the export feature to github but they didn't π
Hi Mr @mrfakename can you please do something about data export on your site? Requested HF to at least upload the export feature to github but they didn't π
Do you mean allowing users to import data into https://chat.mrfake.name ?
Here is the current working method for exporting your data:
- Log in to HuggingChat at https://huggingface.co/chat/login. You will be redirected to the "closed" page after logging in. If you are redirected immediately without seeing a login form, you should already be logged in.
- After logging in, click this link to start the download: https://huggingface.co/chat/api/v2/export. The download is limited to once per hour. Even if you cancel it, you'll have to wait one hour for the next attempt.
If it's not working, please try using a private window (or Incognito tab) in your browser.
It doesnβt seem to be working for me. Iβm not prompted to login and when I paste the export link in there is no indication of a download
Oh no... I can confirm that the method I posted earlier was working just half a day ago, but I've checked again now, and it's no longer working. It just redirects to the "closed" page. This might really be the end. βΉοΈ
What on earth is HuggingFace doing?
Noooooo, just knowing this day, late for 15 daysss
i just found out about this, why would you only save for two weeks, what about people that are on vacation?!?!?! is there any way to back it up still????
Hi Mr @mrfakename can you please do something about data export on your site? Requested HF to at least upload the export feature to github but they didn't π
Do you mean allowing users to import data into https://chat.mrfake.name ?
@mrfakename
While that'd be great too, I meant exporting the chats from https://chat.mrfake.name & download to preserve the chats. Tried using the download link format/api/v2/export but it says not logged in.
Edit: I suppose the export feature was already available for exporting but was kept unlisted on HF. But please update us if that'd be possible. You could open a discussion thread here for your site for everything. & Thank you for your site!
Je crois qu'il faut se rendre Γ l'Γ©vidence maintenant: HuggingChat a fermΓ© et ne rΓ©ouvrira pas: on a perdu tous nos chats sans espoir des les retrouver, c'est fichu ππ’
I missed the announcement before this, if there was one. Disappointed as I had some roleplaying resources I was working on, think I saved most of it but whatever. Nothing world ending. Like the sounds of the new app (or whatever) so will watch for it. Thank you for your efforts!
waiting for whatever you are cooking. and hoping that you will give use more fredom of open source. :). kleep up the good work
Perquisition de la police sur les serveur huggingchat
Sad to see HuggingChat go offline π β it was one of the best ways I explored alternatives to ChatGPT and discovered newer open-source models like Mistral, LLaMA, Cohere, Hermes, Qwen, and more.
Luckily, many of these models are now available through self-hosted or web-based AI chat platforms, so the open-source AI community is still thriving. π±
I hope this doesnβt impact the visibility of newer models HuggingChat helped showcase. Fingers crossed for its return someday β it played a big part in making powerful AI accessible to all. π€π¬
I was working on some content for a tabletop roleplaying game I run. I think I backed it all up... but not sure. Also made an assistant.
Been swamped with work and missed the notice. Hope we get a way to retrieve data but lack of response makes me feel I am SOL. D'oh!
Todos los asistentes desaparecidos y no utilizables?
I was working on some content for a tabletop roleplaying game I run. I think I backed it all up... but not sure. Also made an assistant.
Been swamped with work and missed the notice. Hope we get a way to retrieve data but lack of response makes me feel I am SOL. D'oh!
Unfortunately, the method I posted a few days ago has been disabled by HuggingFace, so I don't think there are any ways left to export the data.
I don't see an option to reply to a specific post, so...
@hugginging5454 I saw that. Thank you for your efforts though!
@Anxosr I mean, I liked mine and some of the others I tried!
I am very sad now. I'm going for a really disappointed walk outside.
Aaaand now the link to huggingchat, from the homepage, returns 502: bad gateway. What on earth is happening?
I propose converting the existing chat interface into an openβstandardsβcompliant generativeβAI agent runtime that is realised as a WebAssembly (Wasm) container executed entirely within the endβuserβs browser.
Inside each Wasm container resides a custom, βunchainedβ software agent that performs all I/O, compilation, testβexecution and UI rendering locallyβthereby supplying the agentβs actuator layer (βmusclesβ).
Conversely, the agentβs cognitive layer (βbrainsβ) is delegated to stateβofβtheβart largeβlanguageβmodel endpoints provided through Hugging Faceβs Inference Providers platform. This separation of concernsβremote inference for reasoning, local Wasm for actionβdelivers lowβlatency interactions, strong security isolation, and vendorβagnostic portability, while remaining fully aligned with the W3C WebAssembly Core Specification and the OCI 1.1 artifact format for distributing Wasm components.
Well this kinda blows π catch me on Venice.ai now I guess
Any desperate last chance ability to export data? I'm so confused. Feels so quick.
hope this will be one of the last best alternatives to huggingchat :((( for now
P/s: ive just created some AI agents, like google gemini or qwen, ............... just in case if anything aint work for me.........
https://talk.shapes.inc/gemini-ai-ggle/dm
https://talk.shapes.inc/qwen-chat/dm
While that'd be great too, I meant exporting the hats from https://chat.mrfake.name & download to preserve the chats. Tried using the download link format/api/v2/export but it says not logged in.
Edit: I suppose the export feature was already available for exporting but was kept unlisted on HF. But please update us if that'd be possible. You could open a discussion thread here for your site for everything. & Thank you for your site!
Any update mr @mrfakename ?
NOOO I NEED HUGGING CHAT πππππππ
hope this will be one of the last best alternatives to huggingchat :((( for now
P/s: ive just created some AI agents, like google gemini or qwen, ............... just in case if anything aint work for me.........
https://talk.shapes.inc/gemini-ai-ggle/dm
https://talk.shapes.inc/qwen-chat/dm
we hope this will be our last bet for now. we don't want huggingchat to die out.
Any Chinese host is gonna be malware. Doesn't mean it's gonna activate right now. But have no expectation that the people who brought you tiananmen square and the great firewall are gonna do anything other than lure trap and kill your data and anything else valuable is unrealistic given the past 2 decades of continued malfeasance.
No como windows
Noooooo. I hope that will not be long
Well, for now, I moved to Mistral.ai Chat Pro + API, and the Groq.com Pro API. The HF Chat Pro was a real deal because one could use and test state-of-the-art models. Now, I'm chatting with Mistral Le Chat Pro, (deepseek, oo.ai, perplexity, basic plan), and my self-hosted Open WebUI connected to local ollama and different APIs. I'm just missing a space where to develop, share, and deploy community models. The HF spaces are very costly if you want to deploy even small models. For the moment, I will pause my Pro subscription because, for me the HF Chat was a really valuable tool, and the other services do not justify the Pro cost for me. What is the best alternative for developing, sharing, and deploying AI models, ideally FOSS; with a fair cost-benefit? Replicate.ai?
Sorry if I missed it but could someone official tell us if there will be an option to export old HF Chat data or was it nuked?
Lack of information/responses kind of disappointing.
ΡΡΠΎ ΡΠ°ΠΊΠΎΠ΅ Π°Π²ΡΠΎΡΡΠΊΠΈΠΉ ΡΡΡ
Waking up one day, and seeing a barely functional export button is mad work!
https://librechat-librechat.hf.space/c/new
I add HF API key and error Something went wrong. Here's the specific error message we encountered: An error occurred while processing the request: 404 Not Found Troubleshooting URL: https://js.langchain.com/docs/troubleshooting/errors/MODEL_NOT_FOUND
Why it doesn't work, and how to fix it
Did everyone miss the fact that they were going to delete your data after two weeks? Did I hallucinate that? If I didn't, your data's gone, folks. Sorry.
Froce Γ vous et Γ l'Γ©quipe π
Hi, quick update regarding personal data, we decided to erase all the user data stored in the database in two weeks. So heads up you have two weeks to back up your data if you want to :)
If you have any doubts about what the data is used for, let me refer you to the privacy policy of HuggingChat : https://huggingface.co/chat/privacy
We endorse Privacy by Design. As such, your conversations are private to you and will not be shared with anyone, including model authors, for any purpose, including for research or model training purposes.
Wish there was a notification for this, I was away and couldn't export...