FREE: Perplexity AI Alternative Perplexica🤖 Open-Source AI-Powered Search Engine 🔍 Llama 3.1 405B

[Music] what's going on everybody Josh pook here and in today's video we're going to be looking at a free opsource alternative to perplexity AI if you're familiar with perplexity AI it is an AI tool that you can use and it's essentially a wrapper um that leverages open AI anthropic other different models to do um Advanced searching of the web and it's very useful but one it's not open source two it's about 20 to 30 bucks a month so we're going to

Show you how you can locally set this up and use it for any model from Gro open AI anthropic olama any open source model you want we're going to show you how to set this up let's dive right into [Music] it all right guys so the tool that we're going to be using that is an alternative to perplexity is called perplex and it's an AI powered search engine and just like the name perplex perplexity it's essentially somewhat of a clone it's

Not a full clone obviously of perplexity but you'll see the UI in just a second and it is very very similar right that's my first thoughts when I saw this I mean you get it when you hear the name it kind of makes sense but um first off before like I'll I'll go through all the links that I'm going to mention here are in the description down below but um you can read up I'm not going to go through their whole

GitHub here but it really just kind of dives into what the features are and if you're using perplexity now which personally I am I'm still paying for it I've been paying for it um you'll you're probably familiar with a lot of these features that I'm going to show you because it's literally just they're trying to clone perplexity here but uh we got co-pilot mode which is in development normal mode um so boost search by generating different queries to find more search

Relevant internet sources like normal search instead of just using the context by SE uh C uh C and NG so that's like an open- Source searching uh tool kind of like an alternative to Google and we'll get into that in just a second how the architecture of this tool per plexa is built but then there's normal mode Focus mode all all mode uh writing assistant mode academic mode so searching like through academic papers YouTube search mode Wolfram Alpha search mode Reddit

Search mode so it's very similar to perplexity like here perplexity we got Focus mode we got uh PR mode AKA co-pilot mode and here we can search like math academic social video this is YouTube math is Wolfram um I believe academic is like all the archive uh papers and whatnot and so it's it's exactly you know really what it they're trying to just clone it um so anyways the installation we'll get into the installation in just one second I just want

To go over here to the architecture just a quickly go over what this tool is actually built on because it's a it's first off it's it's still fairly new um they're they're developing it so they maybe some bugs here and there but um I mean it's been out for I guess like it's been out I guess for a long time in AI years which is I mean obviously ai's been around for a while but in just terms of like the new

Boom of AI and how quickly things develop like there's new developments every single week so it's sort of uh expedited in terms of like time goes by a lot quicker um so yeah it's been around for I don't know exactly when it came around but it's been out for at least this year um a few months ago um maybe even more and anyways so here's the architecture consists of a few Fallen key components user interface of the UI a web base

Interface that allows users to interact with perplex for searching images videos and much more it's built on nextjs um so that's good uh agents chains these components predict perplexa next actions understanding user queries decide whether web search is necessary and then sir x uh C I don't even know sir sirx n uh a metadata search engine used by perplex to search the web for sources and then llms large language models so these are all the large language models like claw gpts

Uh cat GPT open AI as well as you can connect your custom own ones uh and local ones we'll we'll show you that in a second and then embedding models so to improve accuracy of search results embedding models rerank the result using similarity search algorithms such as coine similarity and Dot product distance and then you can read more details there anyways that's a b bit B uh background on the architecture before we get into install you're going to need Docker install

On your system so go to doer. and download this if you don't have it already and if you don't have Docker it's a great tool there's also a very easy way to deploy it if you're running into issu ues and you you can't troubleshoot it and whatnot um I'm going to explain everything in this video um what at least what I believe you could fa potentially face into issues I'm going to cover that so you can get the install right the

First time hopefully but um they link to a tool called repo cloud. or you can just deploy it in a click and uh yeah so if you're having any issues you can check that out it's in the um GitHub right here okay so basically what you're going to do it's very simple so um the do I'm going to cover the docker install you can also install it with mpm I'm not going to cover that but um yeah we're going to install

With dogger so it's very uh simple it's actually a lot easier to do it with Doggers so we're just going to run this git clone GitHub command so you're going to run this in wherever you want this repo to be so let's just go to our handy dandy terminal right here and I've actually already run these commands but I'm just going to walk you through them so once you run that get clone command you're going to go into that folder so

Here's my perplex right perplex right here and you'll see after cloning it navigate to the directory containing the project files rename sample. config dotl to file. config dotl for doer setup you need to fill uh only fill in the following fields and then so yeah you can see my config is just name config it's a Tomo file but you you're just going to want to remove samp all right from that point on you could go into uh the config right here

I'm not going to open it because it has like all my API keys and whatnot but you could add your API Keys like your open AI Gro anthropic Etc but you actually can just change that in the settings too once you dock or compose up now one thing that I I faced an issue which was once I Docker compose up uh then I loaded it it would it was like a continuous loading so the reason being is because it doesn't mention

It just right here so it's kind of like a thing that um that I had to figure out or well they actually have um this wait where is it this section right here explaining it so it's actually in the it's just a bit hidden um in the documentation so you have to go I'm going to leave the link to this as well though just to explain what it is but if you're basically if you already Docker composed up then you can

Just Docker compos down and uh- RMI all to remove the existing per uh perplex perplex containers and images but um you're going to want to open the docker compose right here okay and the docker compose ammo file and basically all you're going to have to do is in the section right here you can see where it says replace IP of 127.0.0.1 with the IP address of the server perplex perplex is running on in these two lines so this is like what

It looks like before you replace that and you can see I replaced it with um Local Host right so I'm just using my local host and you can see here this is the new um next public API as well as next publ uh public WS URL is Local Host for 3001 we have yeah the docker compos uh yl you can check that out and if you need to make any other changes configurations with a port or anything if those ports are

Taken you could do that but other than that once you're' done that you're going to save the docker compos the ammo and then you're just going to rebuild and restart perlex containers so Docker compose up- d double Das build all right so I would just run this command so really like you don't have to uninstall the containers well I don't know if this is something that uh I don't know cuz I I believe some people weren't experiencing this but you know

You may have to uh this is something you may have to do um if you do this then you won't experience the issue that I was facing so uh you could just try and doing it right from the start Docker compose up uh- D and see if you get the issue if you do then you can remove this and then just change this or you could just do this from the start and you should be good to go okay there's also

A bit of a different situation for mac and Linux so they cover that down there but like I said this is going to be linked down below as well as all the other links I've talked about um which is just the other GitHub and whatnot so also too guys you're going to need to install well if you're going to want to run local models you're going to need to install oama if you haven't already and uh once you install oama you

Can pretty much uh download whatever obviously whatever local model you want to run so whether it's quen 2 deep SE coder uh llama 3.1 the new meta model so unless you have like a crazy machine you're going to pretty much probably only going to be use able to use the 8B model um so you just do o llama run llama 3.1 once you actually install o llama okay and then I'm still going to show you how that you can can run

The 70b model and the 405b model as well in just a moment but one thing too just because um the perplex uses an embedding uh embedding as well uh one thing you're one command you're going to want to run you're going to want to run o llama pull nomic Das ed- text okay that's going to be useful if you're when you're using perplex and you're using your o llama and you still want to have a llama embedding um we're going to

Show you that in just a second okay so once you actually get the install going if like I said if you did are experiencing issues or you just do this from the start use um the whole uh exposing perplexed to a network make sure it's your IP okay then you're going to have it loaded up like this okay so as you can see uh this compared to perplexity it's not like 100 like this says where knowledge B begins a perplex says

Research begins here so it's like complete uh replica um it's not like to be quite honest like it's probably not as nice user interface um some maybe you like it better I don't know like it is I do like the black I just kind of like some of this um but some I don't know let me know what you think compared what do you like better for the UI do you like perplexa or perplexity let me know in the comments down

Below but you'll see it's very similar we got like our history here we got search here or so Library we got uh search um now we also have settings here so we'll get into that in just a second and then like I was saying before we have C- pilot and we also have all academic writing Wolfram Alpha YouTube and Reddit so we have all those search options and I'm going to show you a comparison because I have perplexity Pro I've been

Paying 20 bucks a month for it for the last I don't know few months just because I don't know I just I'm paying for a bunch of AI tools right some people you know whatever so I I have been using it and I do still use it a lot but um I'm going to be trying out perplex a little bit more and kind of seeing over the course what I prefer for results like what I think what I think I could

Um yeah like if there's certain situations when one is stronger or whatever the case may be but anyways so um what you're going to do is you're going to go into your settings here okay and I'm going to show you how you can set perplexa up with perplex up with pretty much any model you want so um if you uh basically so here you'll see like olama Gro API key anthropic and these are like AP one these are API keys I'm

Just using for the video and they're cut off anyways so and I'm going to just for safety whatever I'm going to delete them after but um these are the yeah so you can put your API key here so like open AI API key o llama uh so the O llama API URL is um host. doer. internal and then colon 1143 4 so that probably already will be there um but yeah if you need to put that there you can put that

There um I've seen some people say that depending on like the system and this and that some people's are is Local Host uh like this Local Host so if it's not working for you try that but it most people will probably be this and yeah so you can put your API Keys here anytime you change something you can just press this to refresh you have your embedding models here so like as you can see I selected a llama and the nomic

Embed text is right here um for the embed model um and then of course we have dark and light theme obviously we're going to be using Dark theme because we're not psycho I'm joking but um anyways so chat model provider so you could select open AI for example and then over here you could select like turbo gpg 3.5 Turbo gbg4 gbg4 Turbo Omni Omni mini you have those options um and then you could select open AI embedding and do like whatever

Embedding or you could obviously select Gro okay with Gro you have a bunch of options like you could do llama 3.18 B and you're going to get way faster inference time or you could do 70b right Gro currently doesn't have the 4B um 405b right so um yeah but you do also have mistol Gemma Etc so there's Gro and then there's obviously anthropic so Claud Sonet 3.5 Sonet uh really good for coding Etc um and then AMA so with the um

Oh no sorry so ama if you do a llama yeah as long as you just put that URL here then you should be good but the one I'm going to show you to get the 405b model I've covered this in other videos like when I showed I did a video showing you guys how to get 405b within continue dodev on your vs code so if you want to check that out too if you're like want to use that for coding check

Out one uh you can just search continue dodev llama 3.5 or 3.1 on my channel you'll find it basically we're going to be using together. a to do this so you can just sign up for together. I'll leave a link down below and essentially once you do that you're just going to go um log in and you're going to want to go to your API keys and you're going like your settings API key same as you grab any other API key

You're going to want to grab that API key and then you're going to want to come over to custom open AI all right and then you're going to want to paste in I'll leave this down below too so you're going to paste in the model number or the model name which is this okay meta- Lama dmeta or for SL meta llama 3.1 8 B um instruct turbo and then you're also going to want to use this custom open AI base URL

As so API do together doxyz slv1 and then your custom open AI AKA together. a uh API key and then you just update it right so all right guys so we're going to do a couple searches now just to test it out so I'm going to use GPT 40 Omni all right with the text embedding and everything and then I'm going to use GPT 4 Omni in perlex as well as perplex perplexity all right so as you can see I just

Searched latest a I news and we'll go over this real quick um we can see here we got some here are some of the latest AI developments and news in the field of artificial intelligence comprehensive coverage of AI several platforms including CNA okay I'm not going to read through the whole thing but you can kind of see here industry research and Trends uh daily updates websites are dedicated machine learning advancements in various Fields Ai and Robotics computer vision Global AI news

So that's the results there and we can see like the sources here we can also search images and search videos just like how you can in perplexity so very useful um very like you can even see it gives you related questions you could rewrite Copy Text to Speech expand up on it whatever let's do the same search in perplexity so we're going to do here now one thing like perplexities Pro search um Okay so so like Ai and and scientific research

Um government actions of AI corporate so giving like somewhat similar type of response I mean it's a little bit different um I haven't gone through each of them in depth so I can't fully say which one's better um perplexity definitely still does have a bit more features and what not like I would assume it's a little bit better quality I mean I would at least right maybe even more but um it's definitely uh perplexity definitely has a lot of potential I'm

Excited about it and it's an extremely useful tool and it's 100% free open source so you really have no harm using it especially if you're using it with local models AMA like it's definitely uh can be very powerful also too if you don't want to pay for perplexity I mean perplexity look like they do have a free plan at the end of the day so the free plan actually isn't that bad but um you can't use it with higher quality models

The the one thing about perplexity is I do still like like if you use their Pro search and you ask it like a complex question with like multiple different steps it will do multiple queries on it and it's pretty good at that so I mean yeah I'm still going to use both at the end of the day but um this is definitely if I wasn't going to pay for perplexity 100% then I would just use this right and I'm still going

To use both cuz I think this perplex perplex is a really good tool but other than that guys if you're new to the channel uh we upload videos every single day on AI automation AI news Marketing sales things to grow your business with ai ai is exploding and just like every new week there's something new every new day almost it's getting crazy so um if you want to stay up to date with that like this is I'm dedicating my whole business

Life too um you know really trying to master this stuff so I want to keep you guys informed with everything that I learning on a day-to-day basis because I'm always learning new things so I'm going to be put posting these videos for you guys so if you got some value here make sure to subscribe to stay a part of the journey and also too guys if you're new to the channel you haven't already joined our free community we have stric community.com

We going to be doing more free resources in there Etc behind the scenes stuff but if you want to join there we have about 750 to 800 members U maybe even more now in our free Facebook group and Discord Channel combined so you can join there 100% free Link in the description down below you can network with my S as well as other like-minded AI enthusiasts entrepreneurs developers uh marketers business people uh sales people whatever the case may be also too

Guys if you haven't already checked out my book client acquisition code I have a full chapter in here on prompt engineering and this book is extremely valuable if you need uh clients for your business it gives you the ins and outs of everything from cold email uh running paid traffic paid media Facebook ads YouTube ads Etc definitely check that out if you're a business owner or if you're marketer or whatever the case may be salesperson even um a lot of cool

Gems in there other than that guys I will see you in tomorrow's video let me know what your thoughts are in the uh comments down below about perplexa about perplexity and anyways guys keep hustling keep grinding and of course guys accelerate your stride take care