How to Use NVIDIA ChatRTX | AI Chatbot Using Your Files

In this step-by-step tutorial, we dive deep into NVIDIA ChatRTX, an exciting AI tool from NVIDIA. Whether you're a beginner or an experienced user looking to refine your skills, this video covers everything you need to get started. We'll look at how to install and configure various AI models. Plus, we explore how to effectively prompt the AI for both text and image outputs using CLIP.
Host: Kevin Stratvert
📚 RESOURCES
- Get NVIDIA ChatRTX: www.nvidia.com/en-us/ai-on-rt...
⌚ TIMESTAMPS
00:00 Introduction
01:28 Get ChatRTX
02:15 Main interface
02:55 Select AI model
05:05 Select dataset
05:45 AI default model
06:33 Folder path dataset
08:24 Prompt Mistral model
11:46 Prompt CLIP OpenAI model
13:11 Shutdown
13:43 Wrap up
📺 RELATED VIDEOS
- Playlist with all my videos on NVIDIA: • NVIDIA
📩 NEWSLETTER
- Get the latest high-quality tutorial and tips and tricks videos emailed to your inbox each week: kevinstratvert.com/newsletter/
🔽 CONNECT WITH ME
- Official website: www.kevinstratvert.com
- LinkedIn: / kevinstratvert
- Discord: bit.ly/KevinStratvertDiscord
- Twitter: / kevstrat
- Facebook: / kevin-stratvert-101912...
- TikTok: / kevinstratvert
- Instagram: / kevinstratvert
🎁 TOOLS AND DISCOUNTS
✅ 🎙️ Voicemod AI Voice Changer | 5% off | link.xsolla.com/KZBi89AY
✅ 🌐 Squarespace Websites | squarespace.syuh.net/XYaqYM
✅ 🔍 Grammarly | grammarly.go2cloud.org/SH3nL
✅ 🛍️ Shopify | shopify.pxf.io/XY9rPa
✅ 📋 Notion | affiliate.notion.so/rffva4tr71ax
✅ 🖼️ Figma | psxid.figma.com/lqjg97licpry
✅ 🤖 ElevenLabs Text-to-Speech | try.elevenlabs.io/taqepq60mptr
🎒 MY COURSES
- Go from Excel novice to data analysis ninja in just 2 hours: kevinstratvert.thinkific.com/
🙏 REQUEST VIDEOS
forms.gle/BDrTNUoxheEoMLGt5
🔔 SUBSCRIBE ON KZbin
kzbin.info?...
🙌 SUPPORT THE CHANNEL
- Hit the THANKS button in any video!
- Amazon affiliate link: amzn.to/3kCP2yz
⚖ DISCLOSURE
Some links are affiliate links. Purchasing through these links gives me a small commission to support videos on this channel. The price to you is the same.
#stratvert #nvidia #chatrtx

[@KevinStratvert]

Get the latest high-quality tutorial and tips and tricks videos emailed to your inbox each week: kevinstratvert.com/newsletter/

[@AdamsTaiwan]

In 2022 I got a system with a good price for RTX3050 ITX Aero 8G. Never thought it would work well with this app. Thank you for this video. I hope someday it can read all my years of OneNote notes.

[@ThePortraitArt]

Does it support subfolders in dataset and how the size of the dataset relates to CPU and Rem usage.

[@andre1987eph]

play at 2x speed, he's milking it

[@jryde421]

Thank you

[@noobicorn_gamer]

Someone give this more likes

[@rothauspils123]

1.5x speed almost sounded normal

[@JulioBarragan]

is he real?

[@RafaPolit]

I was going to write the same thing: it seems this could have been a 3 minute video with the same information.

[@fifamobile__23]

Unfortunately, my journey ended at system requirements😅. I'll catch y'all on the next one

[@stylie473joker5]

Typical Nvidea lol locking software from old cards

[@TechbaseConsultantServices]

same here ):

[@fifamobile__23]

@TechbaseConsultantServices quite unfortunate 😕

[@luizcamillo9933]

Same here

[@CaptBirdBrain]

Exactly. Me too. We just got Copilot at work and it's pretty good but since it is hosted outside of our local PCs, you have to point it to certain documents. It doesn't actually learn your data. Which from an enterprise security standpoint, makes sense, but this sounded promising because it runs directly from your computer. Unfortunately, neither my work or personal PCs has Nvidia GPUs. At least I didn't waste my time watching this video. One cool feature of copilot, if you have a microsoft account, you can use the Copilot in the edge browser to summarize YT videos (assuming they have a transcript) without watching them. It will even give you timestamps if you want to jump to a certain section. That alone is a huge time saver.

[@Justakatto]

ok this is huge imagine how useful this would be for studying???? like, you can just feed all your notes to the AI and ask it questions and have it ask you questions as your own AI study partner!! OH and you could feed it like the past 5 years' of exam papers to basically deduce which areas of the curriculum that you need to focus on more and stuff like that that's amazing.. ground breaking even

[@bravexcallofduty7629]

hey dumkump ikr this is an awesome news can yu give me more ideas and ways how i can use this to help me in my studiies, thanks

[@ScottWinterringer]

useful... we have had this for over a year... and its opensource...

[@hope42]

💯Important: how to present system requirements when demonstrating a local AI model. Kevin nails it and everybody should do exactly what he does. He saved me so much time since I don't even have that series of Nvidia. He also presents CPU, memory, VRAM requirements which is so important to to helping us geeks go down a rabbit hole we shouldn't have been in. Anyone been there? All AI KZbin channels showing how to run local models need to copy Kevin, in the style of Kevin haha. Mr step by step! The most important step upfront in the beginning of steps is system requirements. Si? 😮

[@ScottWinterringer]

or dont use this... as this is a junk vs the opensource community

[@youMEtubeUK]

Used it a few times, have a 4090, but it’s limited on the number of document and size. So still use Gemini 1.5 pro for 1m tokens books or a bunch of 25 pdfs where it could list and summarise all 25. Local RTX couldn’t.

[@pedrobarros2554]

Whats rtx limits?

[@diverstory]

Wow, amazing stuff! Thank you so much

[@michaellavelle7354]

Excellent presentation. Thank you.

[@KevinStratvert]

Glad you enjoyed it!

[@shinn-tyanwu4155]

Excellent presentation thanks 😊

[@fslurrehman]

Please explain how to use RAG with phi-3, and llama locally using a PC. Also how to install phi-3 on mobile and use it locally.

[@1enonly]

Thankyou kevin !

[@vladyslavkorenyak872]

This is a good improvement. I hope they don't abandon it and keep developing it. Fine-tuning, multiple agents, bigger models, etc. If it could parse through a lot of data and reference with more precision, it would be a wonderful assistant.

[@KevinStratvert]

Agreed!

[@ScottWinterringer]

why? we have had this for a long time.... localGPT

[@joshteckcompany]

This is so nice

[@Shift-consulting]

Thanks a lot for showing that tool. Is it working with sub-folders and the files there as well, or just with the files in that specific folder?

[@claytonhouser7578]

It works in sub folders. But only basix text and pdf folders

[@ThePortraitArt]

Does it support subfolders in dataset and how the size of the dataset relates to CPU and Rem usage.

[@MapedMod]

try it out. it is free

[@streamdungeon5166]

Creatives are usually hating on AI. Now imagine how useful this is going to be for writers referencing across a huge amount of world building materials! I will use it that way and I am already salivating!

[@musicmane4146]

Stealing others content 😅

[@streamdungeon5166]

@@musicmane4146 Did you post in the wrong place? I fail to see the context. I was talking about referencing my own world building text. Sometimes it's better to read before you respond 😂

[@PawFromTheBroons]

Having this query and analyze XLS* files would be invaluable.

[@KevinStratvert]

Then it'll be like the ChatGPT data analyst, which would be amazing!

[@ScottWinterringer]

@@KevinStratvert localGPT like this..... which we have long had...

[@user-uw9ir7fl8l]

i may be hoping too much from my rtx 3050 8gblol all the best kevin

[@Kumodot]

Is it possible to add Metadata to the photos also (inject on exif) with the tags generated by the AI ? I was trying to do someting like that with ComfyUi

[@agua9999]

KEVIN‼️ i subscribed after watching your awesome SQL tutorial. Can you please do a TABLEAU for beginners in the same style? THANK YOU❣️ 🙏🏻

[@bamh1re318]

great intro to RTXChat! Now I have to pony up for a computer with RTX 3060 or 4060,

[@KevinStratvert]

Make sure it also includes at least 8 GB of GPU VRAM.

[@minimal2224]

Guys…I think Kevin is a GPT…his content is to well put together with clear direction. A human could never 🤖

[@KevinStratvert]

haha, maybe in the future I'll be able to trick our AI overlords to keep me around.

[@alexandrunistoroiu452]

Guess it works anyway regardless the gpu! Did you try? Wonder if there will be a linux version...

[@internetnickname8923]

Is it worth downloading other language models if I need to analyze data and afterwards write code in python or will mistral handle it?

[@alitalalhaidar6488]

What's the maximum amount of data present in a folder it can handle?

[@mainasahi7935]

You are brill. any chance you could do a tutorial on Substack

[@rosemaryng7994]

So this is quite similar to Microsoft copilot. But cheaper upfront?

[@alitalalhaidar6488]

Can you have more than one folder path in the same session?

[@herolooter3868]

Which one i download 35gb or 12gb version?

[@MWulffy]

Does it only work in other languages than english ?

[@Rabid-Pinocchio]

First, GREAT info in your vids. Really useful stuff so thank you. I'm a little off topic but I wanted to find a recent vid. I found one of your instructional vids from about a Year ago re. WhisperAI. fyi Whisper now works with Python 3.11 but not 3.12. In your video you prompt a "-U" command to update but that doesn't seem to do anything so I'm guessing that I'm up-to-date. I've seen some referance to Whisper alternatives that are "faster", "more accurate" etc. Since Whisper seems to be working, more-or-less fine I am of the 'if it a'int broke, don't fix it' ilk. That said, do you have any more recent data re Whisper than the >1Yr. old information? Thanks in advance.

[@relexelumna5360]

Why i dont see Add new models, even though i already install and its 32.7gb size? How do i get that?

[@twilightfilms9436]

Can you ask questions in English related to documents, books and papers you have in other languages like French, German, and old languages like German gothic? I’m asking this as a tool for historians and researchers……thanks in advance.

[@amlaaaa479]

Will this become available for iOS?

[@john_blues]

I installed this today. Early impressions(I may update this if I don't decide to delete it): I am using CLIP to search images. RTX 4090. *It does scan subfolders. *It sucks at truthfulness/identification. I asked it to show an image of a lion. I had an image in a sub-sub folder. It show the image, but also 2 images of people. 😵‍💫 *'Show me an image of a car'. Showed 1 image of a person in front of cars, and a folder full of images that had nothing to do with cars. *It never seems to return a null result. So, it's very hit or miss. Which for me means I can't use it for anything useful yet.

[@john_blues]

Using Mixtral: Can you read any files in this folder? I'd love to, but I'm afraid I can't. You see, I'm a text-based AI model, and I don't have the ability to access or read files on your computer. I'm here to help you with text-based tasks, not to browse your files. lol. It did answer my questions. It's just funny that when I asked it directly, it gave me that answer.

[@skistenl6566]

It seems like there are many people like me can't even install it because of this error "An error occured while creating MiniConda environment." NVIDIA's announcement just says keep the default installation path, which I did but got same error. Does anyone know how to fix it?

[@tanteju944]

Does this "AI" work also as face recognition? I would like to create a list of al photographed persons on my thousand of fotos and videos.

[@john_blues]

The CLIP model can, sort of. It's very hit or miss. I assume you mean can you have it find all photos that contain faces, not specific faces.

[@eraser465]

How do you go about getting a copy of the CLIP model?

[@KevinStratvert]

You should be able to click on the drop down to add models and you'll see CLIP there.

[@mydetlef]

And next time when I ask Mistral on the Net it shows me the same numbers and files?

[@virtual_space399]

Sir any math solving problem free ai..

[@josephchoo1153]

just wondering if this chat AI able to read epub files?

[@studiophantomanimation]

This sounds like it would be useful for writing a multi book fantasy series

[@KevinStratvert]

Excellent use case!

[@WickedMagic]

I am so confused to why it works so smothely for you. When I ask the same questions it, simpy answers that it deosn't know the information. Even when telling it which file it is, it just proceedes to inform me how I can open the file.

[@SHISUI1185]

What about code projects??

[@james.telfer]

Follow up request: how to install different LLMs. Had a look on Huggingface but haven't worked out how to install those. Also ChatRTX doesn't seem to reference all items in a folder when answering, just the first one that seems to have an answer. So if you need a summary of multiple files, it won't do it.

[@peteratkin3788]

I suspect its a deliberate limitation, noticed this my self was hoping to use this to scan a client directory, with such a limitation this has limited use case, or maybe your expect to program such things your self, maybe time to hire a programmer.

[@user-yi2mo9km2s]

Its installer has problems since v0.1, Nvidia never fixed until now. Mine is 4090.

[@pamesa]

Hi Kevin. All my files are on One drive. When I tried to use Chatrtx it start to download to my computer all the files

[@KevinStratvert]

That's expected. It needs a local copy of the files first. If you keep your files in the cloud only and don't also have them synced on your PC, it'll need to download them first.

[@JonNewlyn]

This looks like it would be great, but my install failed. I realise it's 0.3 version, but did you have any issues installing?

[@james.telfer]

I noticed large spikes in RAM and swapfile use just as it was completing the install. Try ensuring you've for plenty free. It failed on mine first couple of tries as I had swapfile set to only 16GB

[@JonNewlyn]

Great it worked!

[@BibhatsuKuiri]

i dont have "add new model" tab.. how to get that ?

[@KevinStratvert]

Did you download and install today? The latest bits landed this morning.

[@anthonytsang1360]

Hey Kevin, i have an asus laptop with nvidia rtx 3060 with 6gb of ram. Upon installing i get an error message saying that the chatrtx requires a minimum of 8gb of gpu memory. Any solution to this?

[@KevinStratvert]

Unfortunately, not. You need a min of 8GB of GPU VRAM.

[@saranglee7011]

Heated up my 4090 more then me playing Crysis 4 all maxed out.

[@julianekassler4970]

I uninstalled and download it again to make a clean install. I find it annoying that there are almost no infos about the update that lets you add your own models or where you see clip. It doesnt show up and nVidia doesnt even maintain a change log list.

[@irprocess]

Just tried the software, but it doesn't work as expected

[@ssmd7449]

Does AMD have an equivalent 😅

[@Rehan_Production]

❤❤❤❤❤❤❤

[@umairanees2613]

Might be huge for researcher and scholars. I have hundreds of research articles in my PC and no convenient way to extract required information.

[@KevinStratvert]

That's a great use case!

[@justsurfin5013]

So, NotebookLM?

[@iiFr3on]

can it help you with coding?

[@allHeueldingse]

i want an on cpu or maths copresasor wirh a sac for emulating rtx and cuda over òver gl or gpu's'asm

[@chriscanhigh2991]

an error has occured while creating miniconda environment :( How can i solve it please? i need the requirements.

[@learnanddevelopment7453]

Me too Bogus app

[@InspiredDailyStories]

LLM eat alot of GPU so if you don't have it might as well use the free web base bots.

[@vp5335]

Can it read excel and show graph data ?

[@KevinStratvert]

Not currently. XLSX is not yet a support file type, but hopefully soon.

[@Strawberry_ZA]

Installer is still broken - if you want to change the default installation path it becomes unresponsive, TOO BAD!

[@srirammurthy2739]

Please help me on how to train organisation specific python codes, by the way, is the ChatRTX generate code?

[@stylie473joker5]

Does it work with windows 10 ?

[@KevinStratvert]

The system requirements call out Windows 11 but the user guide specifies 10/11. It’s worth a shot 🤔

[@stylie473joker5]

@@KevinStratvert Thanks downloading now !

[@ScottWinterringer]

I have no idea why I would use this over other opensource items that do a lot better job and I can use any model I wish...

[@patrickbuswell]

My consumption of chocolate chips cookies has tripled since I started watching your videos. 🤬🤬

[@joshteckcompany]

How can I create better content?

[@intermixhector2902]

learning listening reply

[@waiyankyaw8395]

Yeah! First comment 🥳

[@cpt.ahab_it7044]

I am hoping they made the 16GB 4060ti with AI in mind.

[@v.w.singer9638]

It looks interesting, but I honestly can't think of any real use for it. For example, with thousands of images on my HD, it would be faster to browse thumbnails by directory than trying to describe something in an image in a way that wouldn't get me hundreds of results. Or looking for a character in a book (again thousands), especially if I can't remember the name in the first place, or the exact situation.

[@SAVONASOTTERRANEASEGRETA]

I honestly expected something much better from NVIDIA. The software is very heavy and above all incredibly slow. And I have a very respectable PC.

[@crazykkid2000]

Bro ai is resource heavy extremely

[@Anthony-dj4nd]

Sucks that they blocked out older card users in order to try and sell they're new hardware 🕵️

[@mikemorrison7774]

SO.... you can only look at ONE folder at a time??? that's pretty much makes it useless.

[@EBackwards]

I'm going to call it now... AI Model X needs 6GB of Memory. AMD/Nvidia says that 6GB Model needs 16GB's! Want another Model that does more, oh that needs 32GB's! (Really needs 10). What better way to force people into buying their new hardware! Just over inflate the requirements for LLM's in order to sell more hardware.

[@lowqchannel]

i want video generation with RTX free software

[@KevinStratvert]

Me too! Although that'll probably require a GPU upgrade 🤣

[@velvetjones8634]

ComfyUI

[@JavierOropeza08]

rip 2080ti, looks like I might need an update

[@graphguy]

Cool but data types is pretty limiting … no excel , no images.

[@graphguy]

Corrected… the current website says: Chat With Your Files ChatRTX supports various file formats, including txt, pdf, doc/docx, jpg, png, gif, and xml.

[@aivy-aigeneratedmusic6370]

Nvidia could create the best system with the best interface. But they don´t. Right now several Github Projects are better, support more file types and have more features.

[@fransoualee6529]

Nvidia: 8GB VRAM minimum requirement... while also telling everyone that their 8GB VRAM GPU are more than enough.

[@KevinStratvert]

🤣

[@solomonali7229]

World go cookies

[@caseyczarnomski8054]

This software is buggy. I downloaded it and during setup I got errors that whisper and the other add on will not install during setup. Trash.

[@garyrichardson475]

I’m still waiting for Kevin’s apology or explanation about his last short in which he was pushing a “get paid for promoting games” scheme. Check out his most recent entry and KZbin Shorts. Until he does that, he’s lost his credibility with me.

[@KevinStratvert]

Hey Gary, thanks for the note and feedback. Check out my community page for more information. Thanks. 🙏

[@ArvindhCM-fu6tv]

Seems ChatRTX didn't get the memo that my G14 packs an RTX 3060. . Oh well, the disappointment is real!

[@kaleidoscopeon]

I stopped watching when the requirements asked first an rtx gpu. No thanks

[@zoom-into-nature]

😂😂Sorry

[@statesman6379]

This is how Microsoft gets your credit card numbers, soc, and passcodes.

[@KevinStratvert]

I don't think I'd store that type of information unencrypted on your PC 🫤. That said, this runs locally without a need to connect to the internet and is unaffiliated with Microsoft.

[@vincentcheung8504]

This AI is basically useless..

[@Alex29196]

Kevin, I've begun to question your content because you endorse this product used by only 5% of users, which has been shown to be far less efficient than any LLM inference interface. Good luck with promoting Nvidia.

[@Futurethinking2337]

So tired of the cookie company examples. Maybe use something else once in a while

[@KevinStratvert]

What kind of business do you recommend instead? I need some type of example 🤣

[@WaqasAli-dq9wg]

Let's try to reconnect for competition 😊

[@WaqasAli-dq9wg]

Hello World of Cookies 🍪🎉😊

[@SlippstersVideos]

NO. I DONT WANT IT.