FREE: Twinny AI Coding Ollama Llama 3.1🤖 Github Copilot, Continue Dev, Claude Dev Alternative

[Music] what's going on everybody Josh poock here and in today's video I'm going to be showing you a 100% free local alternative to GitHub co-pilot no you don't have to spend 20 to 30 40 bucks a month on GitHub co-pilot you can use a tool called 20 all right it's available on GitHub I'm going to show you how you can download it get it set up and use it with any model you want whether that's Gro open AI Claude olama any

Open source model llama 3.1 whatever the case may be let's dive right into [Music] it all right guys so here is 20 links to everything will be in the description down below and I'll give a quick overview on 20 in just a second this is somewhat of uh you you may have you know if you've been on this channel see me cover tools like continue Dev cover tools like claw Dev and bunch of different ones and I'll I'll briefly touch on

Some of them uh in this video just kind of mention how I think 20 compares to them in certain areas this isn't a comparison video but definitely recommend checking out those tools as well and I'll I'll dive into a bit more of those in just a second once we start looking at 20 so here's 20's website as you can see the free and private AI extension for visual studio code so this is one thing I kind of like about it you

Know um we've also covered tools like AER Claud engineer AMA engineer um bunch of different different tools um that you may have to you know set up a little bit or do pip install whatever the case may be this one's really nice just because uh s similar to continue Dev and similar to claw Dev for um claw Dev yes claw Dev and continue Dev this is just a uh Visual Studio code extension so you're using visual studio code or cursor whatever

The case may be you can easily get this up and running so I'll quickly walk you through some of the features here so they have a getting started quick uh quick start guide right here and their docs which pretty much cover all the questions you could have in terms of Quick Start inference provider supported models uh the chat the fill in the-middle which is autocomplete uh keyboard shortcuts uh symmetry Network and all that good stuff so I'll leave a link to

The docs too and here you can see it in action fill in the middle code completion right here um chat with AI about your code so it's very similar to models we've or different tools if you've you know if you've been looking at AI coding tools you probably are familiar with such as continue Dev GitHub co-pilot has those basic features that we want when we're doing coding um so as you can see chatting with the code here operates online or offline

Highly customizable API endpoints for fim and chat uh chat conversations are preserved conforms to the open AI API standard reports single or multi-line fill in the middle completion customizable promp templates generate get commit messages from stage changes um and yeah a bunch of other stuff I'm actually just going to show you how to set it up and then I'll actually show you all these features once we're in there so without Ado let's go ahead and install 20 so you can go

Directly to this link if you want or you could simply just go into your Visual Studio code and go into the extensions right here and just search 20 okay and you'll be able to download that install that once you download that install that one thing you're also going to want to do um especially if you're going to use local models is of course you're going to want to have oama downloaded and installed so I'll leave a link down below for that

You can download o llama and then it once you download o llama depending on which model you would want you would want maybe llama 3.1 or whatever the case may be you're going to go here and just run this command okay one cool thing I'll also mention about 20 is if we go ahead and open up our Visual Studio code um first once you download it it will show up right on the left hand side now I always like to and

I'm going to drag my screen over here just so you can see and I'm going to move my head over to the left hand side just for now um as you can see this is the chat interface so I I always like to drag mine over here and put it up here as you can see I have continue here I have claw Dev and I have 20 now and yeah I just like it it's a little bit slick sleeker for me

Having it on the right hand side now we're going to go over some of the features here and then I'm going to show you how to make sure that you get your own llama and whatever models you want set up first things first um actually I'll just show you right away how to get the Alama model set up so if you go to manage 20 providers up here you'll be able to see um the different they're going to have default models

Um what you can do is click on reset providers now if you're having trouble one thing I did you may or may not need to do this I just went into the settings and there is a lot of different settings here and you can play around with some of these if you want I'm not going to cover them all um and I don't even know if this was necessary but I just did it and I mean it seems to be working

I mean I may have been working without this but initially it was 0.00 whatever I just put it to Local Host but either way actually may work just cuz 0.00 uh is right here too so that's just a little side note you may not have to do that the main thing is you're going to click oh I just messed it up well whatever we'll fix it right now so you're going to click reset providers and then you're going to come in

Here and click on edit so now I'm editing this provider and for me I'm going to choose a llama for the provider so same thing and this is the chat so I'm going to use llama 3.1 latest and then I'm going to save that and then I'm going to come here to the fim and I'm going to put this to llama all right and then I'm going to just cuz I'm using llama and then I'm going to put it to llama

3.1 save and then and then I'm going to come to the embedding right here and I'm going to use my n nomic embed text latest but I mean you can use whatever embedding model you want and this is kind of something a little bit different than some of the other tools we covered such as continue Dev um claw Dev I actually don't I mean I know claw Dev just came out with a few updates I haven't fully checked all them out

Yet so I don't know if they added some other stuff like this but this is something kind of cool I'll show you in just a second but they have uh embedding features so yeah like a rag essentially so now if you were to use like open AI or anthropic or whatever like Gro um whatever the case may be you would just put the host name here as the uh endpoint uh the base URL all right so whatever provider you're looking for

Actually uh yeah you would set up set up the host name sorry this is the host name and then the API path would be the the folling path of that uh URL so like open AI whatever search it up on you can you ask chat GPT what's uh anthropics base URL or open AI whatever and then you would just put the that there without the uh the forward slash uh path all right if it's a different path then just change that

There and then you should be good and then you're going to want to put the API key as well and yeah and um you would want to um maybe change some of this around https and whatnot so anyways so I'm just going to save that and yeah so you can mess around with the providers add different providers Etc once you do that um we can go over to start a new chat and I can actually just go back and say like

Hi actually well first off if you click the AI uh Emoji right here you'll see the different options for your chat so right now I'm using the one that we just selected llama 3.1 for chat and llama 3.1 for fill in the middle um here we have embedding options so you can see the embedding provider right here the chunk Max chunk size the minimum chunk size the overlap you can embed workspace documents code Snippets the number of code Snippets used as

Context file pass the number of file has to be used as context search metrics so um cosine uh Dot and you know the different ones so um and then the metric to be used for the vector search and then rerank probability threshold 0.5 so the lower the threshold the more likely a result is to be included all right so this is kind of cool this is a newer feature I believe that they've gotten which essentially um allows you to uh leverage

Rag Within your coding experience and whatnot um and then here is toggle autoc scroll on and off so basically when you're just chatting with it like if you're having a large chat it's essentially just going to auto scroll on or off here we have generate commit messages from stage changes so it will autogenerate commit uh messages when it's generating your code it can add files in your uh code base and and generate those commit messages so here we're we're seeing 20

Is talking to us right here using l 3.1 100% locally 100% free we're not paying for um GitHub copilot or anything like that here's a scroll to the bottom and then here's to enable SL disable uh rag context for all messages so you can enable that or disable that and then up here we have uh symmetry network settings so you could check some of this out um provider connection uh consumer connection right here so and then we have manage 20 temp

20 templates so right here you can actually create templates um or use these default templates from 20 that will basically um be things that you can leverage like basically prompts for your 20 chat interface so add Tas add types commit messages Etc and yeah so you can create new ones here's where you can can view your 20 conversation history as we can see this chat right here um so this is all kind of cool because it does have some specific features

That some of the other ones that I really like like continued Dev or claw Dev don't have I mean I haven't checked out all the new claw Dev updates but um yeah to my knowledge neither of them have the whole rag functionality um which is kind of interesting and kind of cool um so yeah I know Claud Dev can leverage anthropic new prompt caching and whatnot but um but yeah so so now that we kind of went over everything let's just

Basically um ask it to generate us some code generate something cool things calculator here's a basic python script generation displays 10 random cool things on the screen okay so as you can see it's generating the code you know little bit slow this is just what happens when you're using a uh open source model you know depending on what model it is and depending your Hardware of course okay so once it generates the code we have three options we can accept the

Solution right here we can copy the code and we can append to new document so let's go ahead and test this appending it to a new document and boom as you can see we got a new document right here import random and boom we have uh the code right here and uh I would probably rename this to something a little bit better all right so as you can see we ran Python test.py and it told us 10 random cool things so

Time travel will revolutionize the world time travel will arrive in the year 2027 robo dogs is currently in development robo dogs will arrive in the year 2025 a year from now I mean who knows with AI this intelligent we may see it soon so anyways once you do that you could actually um so you Al also can uh use the um you can basically highlight specific parts of the code and you can right click on it and you'll see here you

Have different options from 20 which is pretty cool you can uh add types explain generate docks refactor or write tests continue Dev also has um some stuff like that too but let's go ahead and explain this right here so we're just saying explain and as you can see and as you can see it generates an explanation of the code so that's just a few things 20 can do um it's a pretty uh useful tool um it's it's hard to say like

How it compares to quad Dev or continue Dev I'm going to continue using different ones and testing them out and I'll keep letting you guys know um what my favorite ones are I really do like continue Dev I really do like claw Dev I know a lot of you guys like them too um 20 is a pretty cool useful tool as well so uh yeah it it does have some capabilities that some of the other ones don't such as the rag

To my knowledge and whatnot so I mean if they continue improving it and updating it and coming out with new updates then it could be um definitely a very you know top-notch tool potentially right um so other than that guys hope you got some value from this video if you did make sure to drop that like and uh hit that subscribe button if you're new to the channel we upload videos every single day on AI automation coding development business growth Marketing

Sales Etc so make sure to smash that like button and subscribe to stay up to date with the daily uploads uh also too guys if you haven't already joined our free community stri community.com our free Facebook group and free Discord Channel link for that will be in the description down below other than that guys I will see let me know what your thoughts are too about like continue Dev 20 claw Dev what's your favorite one um if you have any ones

Specifically that I haven't covered or that you just maybe don't know and you want me to cover or any tool in general regarding AI or a video on a specific thing let me know in the comments down below I know some of you let me know in previous videos I just have a lot of different video I stuff you know that I need to create so don't worry I may um if I see your comment actually create a video on it

So let me know other than that guys I will see you in tomorrow's video keep hustling keep grinding and of course guys accelerate your stride take care