This is Awesome 🤜🤛. I have trailed countless llm chatbot over the past 6 month's with varied results but none have come close with respect to the accuracy = trust. Anything LLM is so close to being a perfect all-rounder i am really looking forward to watching this develop into something super amazing. Well done team Anything LLM. Will be in contact soon with some realworld UI suggestion and features. 🎉
hi, this is really cool. this is really amazing work, thank you. i have a few feedback points, maybe they are on the upgrade list already: 1. add user payment (paypal, stripe, ... etc) to individually price different workspaces or charge monthly subscriptions 2. add user registration 3. allow users to upload documents 4. maybe a front page to customize would be nice to have with the user registration and login 5. add a place to upload a local llm 6. add a place to download the local embedding files and vector db fore each separate workspace 7. all a place to put privacy policy with the links at the bottom left. again thank you very much for the amazing work, this will get huge very fast.
this looks like an amazing project. I am going to try it out now. Would love to see multiple AI agents in single workspace. Something like microsoft's autogen incorporated in. Maybe give each agent a different vector db and again in same single workspace. Like workspace for marketing, and inside have lets say we create a marketing dept workspace and in there we have copywriter agent with its own vectordb related to copywriting, we can have a planner agent with its own marketing planning or strategies vectordb etc. That would be amzing :)
Tim this is super cool have been looking for a self hosting option…as have a bubble SAAS for this…can you hook up langchain to this? Specifically langflow/flowise to extend functionality thx
Hate to sound like an ingrate but until this supports local LLMs I will hold my stoke. Side bar, I will admit I am more of a layman but how is this different from using langchain? Is it just a Ui to the same tech that langchain uses?
I have company knowledge base docs that are probably 50-70 pages of info (all well labeled but disorganized) could AnythingLLM COHERENTLY read the document and organize all of it into an outline so I could turn 50+ pages into well structured and nested info (like a table of contents). I’ve seen many AI “doc readers” make claims but in practice they all miss tons of info, and barely provide any real value. Please tell me this is different haha
Currently, this tool allows you to chat with the docs via referring to context in the provided docs - it doesn't do summarization _yet_ where it can refer to everything. So right now, no, but we also dont make that claim!
@@TimCarambatCould you see this being a feature in the near future and if so how? Would it require an improvement from llm side or is there anything you could do? Thanks
Awesome app. I’ve been trying to figure out multi unique user sessions for a bit now on my chatbot. Do you have any documentation you can point me towards that might help?
impressive software how do you choose a custom path to store the documents you want the workspace to use? Is it a windows env variable, if so what is it?
@TimCarambat Hey Tim, Can you share what is the New Thread that you have on your new version? I understand the new WS, but for the new Thread, will it have a separate DB for the conversation or will it keep track of the memory and data from the previous Thread? why do we need to create a new one? is there any word count limit? tnx
Threads will share the same documents, but have unique histories. The only limit on any given thread or workspace is your models context window - we dont artificially limit it!
Is it possible to select specific text splitters for the uploaded documents? For example to use html delimiters or Java delimiters depending on the file extension? Thank you
Do you happen to know if it is possible to hook up AutoLLM to this? I can see the other local options but they all require embeddings whereas AutoLLM has embeddings built in. Thanks!
Gotcha! The search response is like night and day compared with AutoLLM since it has built in RAG. It would be amazing to see it wrapped into AnythingLLM! 👍@@TimCarambat
Im getting Troubles already 🙄Do I need to pay openAi for the API Key...? Even for ChatGPT 3.5? Or did I setup soeting incorrectly Debuggung Info : Could not respond to message. OpenAI::CreateModeration failed with: Request failed with status code 429 Help Pls!!
Did this get sorted out? I'm having the same problem after multiple installs on different machines and doing everything I can think of (which isn't much, as I'm not a techie)
@@AhmedIbrahim555 The only key required right now is for the LLM side of things, which if you weren't pay via an API you would be paying to run GPUs (which would cost certainly more!). However, we do intend to support local/hosted LLMs outside of the typical API-based ones
I'm literally Gobmacked at the possibilities. I had a meeting yesterday with a lawyer who needs exactly this. Locally installed. I am setting it up for myself now on my pc.
Ah, we are missing a GPT4ALL integration. However we do work with Ollama and LMStudio and LocalAI - all of which can run the same models locally and offline
If you look at the code we have a Mic icon commented out right now - we were going to add it eventually for ease of use. Just not live right now. Best would be STT via your phone or devices native keyboard action
@@TimCarambat I still cannot figure any way to use finetuned local model in windows. Could you please guide me? I do not see any option to select local AI model at all.
This is awesome! Quick question. For a local install is there special configuration needed to upload documents? I got it up and running but the upload appears to be disabled and shows "Document Processor Unavailable". Any help is much appreciated!
Are you running the document process in another terminal (or thread)? Its a python API so the setup is a bit different but it is being worked on to be more simple to run at startup. github.com/Mintplex-Labs/anything-llm/blob/master/collector/README.md#setup
@TimCarambat OK, thanks for info! I wasn't running it as a thread. Do I need to use the Docker implementation to use this method you provided or can I run it in a terminal along with front and backend processes?
I believe we are meant to be like Jesus in our hearts and not in our flesh. But be careful of AI, for it is just our flesh and that is it. It knows only things of the flesh (our fleshly desires) and cannot comprehend things of the spirit such as peace of heart (which comes from obeying God's Word). Whereas we are a spirit and we have a soul but live in the body (in the flesh). When you go to bed it is your flesh that sleeps but your spirit never sleeps (otherwise you have died physically) that is why you have dreams. More so, true love that endures and last is a thing of the heart (when I say 'heart', I mean 'spirit'). But fake love, pretentious love, love with expectations, love for classic reasons, love for material reasons and love for selfish reasons that is a thing of our flesh. In the beginning God said let us make man in our own image, according to our likeness. Take note, God is Spirit and God is Love. As Love He is the source of it. We also know that God is Omnipotent, for He creates out of nothing and He has no beginning and has no end. That means, our love is but a shadow of God's Love. True love looks around to see who is in need of your help, your smile, your possessions, your money, your strength, your quality time. Love forgives and forgets. Love wants for others what it wants for itself. Take note, love works in conjunction with other spiritual forces such as faith and patience. We should let the Word of God be the standard of our lives not AI. If not, God will let us face AI on our own and it will cast the truth down to the ground, it will be the cause of so much destruction like never seen before, it will deceive many and take many captive in order to enslave them into worshipping it and abiding in lawlessness. We can only destroy ourselves but with God all things are possible. God knows us better because He is our Creater and He knows our beginning and our end. Our prove text is taken from the book of John 5:31-44, 2 Thessalonians 2:1-12, Daniel 7-9, Revelation 13-15, Matthew 24-25 and Luke 21. Let us watch and pray... God bless you as you share this message to others.
The API key being used it being Rate-limited. This could happen for many reasons if you are using the same key across multiple scripts or accounts. Otherwise, are you sending a bunch of chats very quickly? 429 is the rate limit error.
Very well done Tim. Nice UI, good features, professional Web site. Keep up the good work!
This is Awesome 🤜🤛. I have trailed countless llm chatbot over the past 6 month's with varied results but none have come close with respect to the accuracy = trust. Anything LLM is so close to being a perfect all-rounder i am really looking forward to watching this develop into something super amazing. Well done team Anything LLM. Will be in contact soon with some realworld UI suggestion and features. 🎉
This needs more press. More people need to know about this. Awesome!
Great job Tim, major accomplishment and well executed OS project. Impressive!
hi, this is really cool. this is really amazing work, thank you.
i have a few feedback points, maybe they are on the upgrade list already:
1. add user payment (paypal, stripe, ... etc) to individually price different workspaces or charge monthly subscriptions
2. add user registration
3. allow users to upload documents
4. maybe a front page to customize would be nice to have with the user registration and login
5. add a place to upload a local llm
6. add a place to download the local embedding files and vector db fore each separate workspace
7. all a place to put privacy policy with the links at the bottom left.
again thank you very much for the amazing work, this will get huge very fast.
this looks like an amazing project. I am going to try it out now. Would love to see multiple AI agents in single workspace. Something like microsoft's autogen incorporated in. Maybe give each agent a different vector db and again in same single workspace. Like workspace for marketing, and inside have lets say we create a marketing dept workspace and in there we have copywriter agent with its own vectordb related to copywriting, we can have a planner agent with its own marketing planning or strategies vectordb etc. That would be amzing :)
was thinking about sth exactly the same
I agree, this feature should be top priority!
Looking great. Seems so much more user friendly that others ive tried. Quivr was a pain. This looks easy, though i havent had time to try it out.
Dang with a few retrofits, and SPR, maybe add .zip and .js files, we could get this going on very large code repos. That should be a demo!
This looks amazing! Thank you so much for your work on this!
congragulations and thank you Tim!
The product looks great. I watched at 1.5 speed and only then the video is normal speed 😂
All I can say is WOW!!!
I only have one thing to say: Amazing!
Tim, would you please provide a quick competitive analysis ? I like what I see, but there’s just so many tools out there that are reading local pdfs.
Gosh this thing is great. Thank you.
This is so amazing!
How to deploy anything -llm on docker for windows pc. Please make a step by step Video with attached all the codes with database ....
This is veeeeery cool, like for real amazing! Do you have self hosted llm's in your roadmap? Like spinning up an ollama API locally and using that?
We do! It be right after anthropic because then we can support hugging face, ollama, replicate, and a ton more in one fell swoop
@@TimCarambat love it! Thank you so much.
Tim this is super cool have been looking for a self hosting option…as have a bubble SAAS for this…can you hook up langchain to this? Specifically langflow/flowise to extend functionality thx
Hate to sound like an ingrate but until this supports local LLMs I will hold my stoke. Side bar, I will admit I am more of a layman but how is this different from using langchain? Is it just a Ui to the same tech that langchain uses?
You're awesome
Thank you
I have company knowledge base docs that are probably 50-70 pages of info (all well labeled but disorganized) could AnythingLLM COHERENTLY read the document and organize all of it into an outline so I could turn 50+ pages into well structured and nested info (like a table of contents). I’ve seen many AI “doc readers” make claims but in practice they all miss tons of info, and barely provide any real value. Please tell me this is different haha
Currently, this tool allows you to chat with the docs via referring to context in the provided docs - it doesn't do summarization _yet_ where it can refer to everything.
So right now, no, but we also dont make that claim!
@@TimCarambatCould you see this being a feature in the near future and if so how? Would it require an improvement from llm side or is there anything you could do? Thanks
Ook this is huge!
Thanks for the video, Tim! Quick question: Will this work with scanned PDFs, or is it only for digital PDFs?
If the scanned PDF's has been through OCR, sure thing. If not, then idk.
digital only right now, OCR is not built into the uploader
Awesome app. I’ve been trying to figure out multi unique user sessions for a bit now on my chatbot. Do you have any documentation you can point me towards that might help?
I have an issue document could not be imbedded? How to resolve this?
so good!
impressive software
how do you choose a custom path to store the documents you want the workspace to use? Is it a windows env variable, if so what is it?
how do you make money - what's your pricing model?
How to adjust prompt to make model act under a specific role?
@TimCarambat Hey Tim, Can you share what is the New Thread that you have on your new version? I understand the new WS, but for the new Thread, will it have a separate DB for the conversation or will it keep track of the memory and data from the previous Thread? why do we need to create a new one? is there any word count limit? tnx
Threads will share the same documents, but have unique histories. The only limit on any given thread or workspace is your models context window - we dont artificially limit it!
Is the number of chats somehow limited by AnythingLLM?
Very nice, how does this compare to Quivr?
Can't copy and paste into this app, some websites are long and trying to type it out is exhausting
how does this compare to Verba?
Is it possible to select specific text splitters for the uploaded documents? For example to use html delimiters or Java delimiters depending on the file extension? Thank you
I'm a great appreciate, Tim
The video is named "v2", but downloadable is v1.5. What am I missing?
Video was made before desktop existed. Desktop version is 1.5
Is this tool capable of multimodality?
Installed, never saw the option for the white labeling or ability to edit the appearance, only custom messages.
You installed the desktop app. All of that is for the self hosted server based version
I'm encountering this error "message.Reason: OpenAI::CreateModeration failed with: Request failed with status code 429" How do I resolve this?
What is the vector dimension used when embedding?
1536, using OpenAI Ada 002
can you please make a video of how to use hugging face models in anything llm.
What issue are you encountering?
Nice, can we use a web browser to connect to it ? thanks
The Docker version, which is demoed here, yes
Is there any way to access the internet? with plugins?
Chef’s kiss 🤌🏾
Do you happen to know if it is possible to hook up AutoLLM to this? I can see the other local options but they all require embeddings whereas AutoLLM has embeddings built in. Thanks!
We have a built in embedded now so no additional setup beyond LLM! Will have to look into auto llm however
Gotcha! The search response is like night and day compared with AutoLLM since it has built in RAG. It would be amazing to see it wrapped into AnythingLLM! 👍@@TimCarambat
Im getting Troubles already 🙄Do I need to pay openAi for the API Key...? Even for ChatGPT 3.5? Or did I setup soeting incorrectly
Debuggung Info :
Could not respond to message.
OpenAI::CreateModeration failed with: Request failed with status code 429
Help Pls!!
Did this get sorted out? I'm having the same problem after multiple installs on different machines and doing everything I can think of (which isn't much, as I'm not a techie)
How to add my own knowledge base?
Just upload some files into a workspace - done!
i connected with pinecone database but it does seem to be responding from my docs , responses seem its coming firectly from openai
Be sure you embed the documents into the workspace you're using!
Anything for no api please 😅❤?
What do you mean? I dont understand the question 👀
@@TimCarambat Sorry I just used your "anything" 🤗 I meant, is there any strategy totally free without linking stuff with payed keys?
@@AhmedIbrahim555 The only key required right now is for the LLM side of things, which if you weren't pay via an API you would be paying to run GPUs (which would cost certainly more!).
However, we do intend to support local/hosted LLMs outside of the typical API-based ones
💙❤️
This is Freaking UnBelievable!
Can you tell us this, how would we connect to language models we have already downloaded for say "GPT4AL"L?
I'm literally Gobmacked at the possibilities. I had a meeting yesterday with a lawyer who needs exactly this. Locally installed. I am setting it up for myself now on my pc.
Ah, we are missing a GPT4ALL integration. However we do work with Ollama and LMStudio and LocalAI - all of which can run the same models locally and offline
@@TimCarambat I used LMStudio and I was setup in literally minutes...however, I would like to copy over the models.
Tim can I add Whisper as Speech-to-text for my query
If you look at the code we have a Mic icon commented out right now - we were going to add it eventually for ease of use.
Just not live right now. Best would be STT via your phone or devices native keyboard action
@@TimCarambat I love it so far but I hope custom models is a priority. For example from huggingface (or) my own models hosted in aws bedrock
Can this integrate into Slack, Microsoft Teams, and Whatsapp?
It has a full api, so you can make whatever plugins you like. I would love a plugins marketplace in the future for devs to extend their instances on
Is it possible to run with own GPT Fine tune model?
Getting merged today github.com/Mintplex-Labs/anything-llm/issues/311
So yes, but only fine tune on 3.5
@@TimCarambat That sounds great! Thanks for your great product!
@@TimCarambat I still cannot figure any way to use finetuned local model in windows. Could you please guide me? I do not see any option to select local AI model at all.
This is awesome! Quick question. For a local install is there special configuration needed to upload documents? I got it up and running but the upload appears to be disabled and shows "Document Processor Unavailable". Any help is much appreciated!
Are you running the document process in another terminal (or thread)? Its a python API so the setup is a bit different but it is being worked on to be more simple to run at startup.
github.com/Mintplex-Labs/anything-llm/blob/master/collector/README.md#setup
@TimCarambat OK, thanks for info! I wasn't running it as a thread. Do I need to use the Docker implementation to use this method you provided or can I run it in a terminal along with front and backend processes?
Got it all figured out! Thanks again for pointing me to the right direction. Keep up the awesome work! @@TimCarambat 👍
I believe we are meant to be like Jesus in our hearts and not in our flesh. But be careful of AI, for it is just our flesh and that is it. It knows only things of the flesh (our fleshly desires) and cannot comprehend things of the spirit such as peace of heart (which comes from obeying God's Word). Whereas we are a spirit and we have a soul but live in the body (in the flesh). When you go to bed it is your flesh that sleeps but your spirit never sleeps (otherwise you have died physically) that is why you have dreams. More so, true love that endures and last is a thing of the heart (when I say 'heart', I mean 'spirit'). But fake love, pretentious love, love with expectations, love for classic reasons, love for material reasons and love for selfish reasons that is a thing of our flesh. In the beginning God said let us make man in our own image, according to our likeness. Take note, God is Spirit and God is Love. As Love He is the source of it. We also know that God is Omnipotent, for He creates out of nothing and He has no beginning and has no end. That means, our love is but a shadow of God's Love. True love looks around to see who is in need of your help, your smile, your possessions, your money, your strength, your quality time. Love forgives and forgets. Love wants for others what it wants for itself. Take note, love works in conjunction with other spiritual forces such as faith and patience. We should let the Word of God be the standard of our lives not AI. If not, God will let us face AI on our own and it will cast the truth down to the ground, it will be the cause of so much destruction like never seen before, it will deceive many and take many captive in order to enslave them into worshipping it and abiding in lawlessness. We can only destroy ourselves but with God all things are possible. God knows us better because He is our Creater and He knows our beginning and our end. Our prove text is taken from the book of John 5:31-44, 2 Thessalonians 2:1-12, Daniel 7-9, Revelation 13-15, Matthew 24-25 and Luke 21. Let us watch and pray... God bless you as you share this message to others.
ok thank you sir, im am careful of ai thanks to you
who gives a shit. Stop pushing your personal choice of religion onto others.
Why use this when u can build your own
sure, we went and built our own and made it open source but if you would like to redo all the work you can build your own from the ground up too
bruh..... guy did us a huge favor here...
Only dark mode, srsly?! 🤮
Cant please everyone 😤
If you like to stare into the sun
I'm encountering this error "message.Reason: OpenAI::CreateModeration failed with: Request failed with status code 429" How do I resolve this?
The API key being used it being Rate-limited. This could happen for many reasons if you are using the same key across multiple scripts or accounts. Otherwise, are you sending a bunch of chats very quickly? 429 is the rate limit error.