Install Open Interpreter in 2 min | The free, open source CODE INTERPRETER!

#ai #codeinterpreter #openai
🔥 Install our A.I. Newsletter in your brain:
natural20.com/
LINKS:
github.com/KillianLucas/open-...
github.com/KillianLucas/open-...
platform.openai.com/account/a...

[@WesRoth]

TROUBLE INSTALLING? Here's an installation and troubleshooting guide: natural20.beehiiv.com/p/open-interpreter-tutorial-installation-use-cases

[@gingerhipster]

Many people still have no idea the state and ease of access for all of this stuff, including in tech. Your videos do a great job of quickly walking through really complex tasks and subjects. Thanks

[@apache937]

Apple will make it so that their model can do anything the user can do in the OS

[@gingerhipster]

@@apache937 the direct integration is nice, but these models will quickly be able to do anything they can parse the API documentation for. 😬

[@WesRoth]

Thanks! yeah, the goal is to make people realize that this is much more accessible that it appears.

[@BergomyLegendre]

Love all your videos Wes, definitely one of my favorite AI creators. Will fasho check this out!

[@ColinTimmins]

Thanks for your videos, they are awesome! I’m glad I found this channel in all the noise created by people milking the AI trend. Much appreciated! I gave up on coding 25 years ago because of my dyslexia and now I am learning and coding things that work. I have an idea in my head and it’s slowly oozing out into code, what a time to be alive! =]

[@WesRoth]

YES! That's great to hear. Keep going with coding, it's seems to be getting easier each week with all the new stuff that's coming out. Open Interpreter seems like the next big leap.

[@sneakygs]

Hope to see these AI coders (especially open access) in action soon! You mentioned previously you might drop a few applied tutorials. Eg, making an audio to text summary for podcasts. Love your channel, approach, way of thinking.

[@WesRoth]

thank you! Yes, I'm still trying to get the tutorials out, too much content ideas lol. But that should be coming within the next few days. I think I have 2 videos, then the AI app tutorial next.

[@zacharyluttrell2286]

Honestly this feels pretty revolutionary to me. Likely not without its risks or downsides, but after having used it for the past few hours it seems incredibly powerful. Definitely helps streamline the development process with LLMs for me so far.

[@jdsguam]

I asked it to create a Desktop Short-Cut on my desktop that opens the terminal and executes the initial command automatically. I even had it create a voice UI for me and it did; but, it's still a work in progress. It scanned my images and deleted duplicates. It created circle icons from Stable Diffusion Photos that I can use that are superior to windows default icons. Cost me under $5 in all and 3-4 hours of useful entertainment!

[@DCinzi]

I have spent the last month and a half of my life, with no starting knowledge in coding whatsoever, to try to build exactly this. 3 windows and one ubuntu reinstall laters we are here.. awesome.

[@apache937]

virtual machines?

[@DCinzi]

@@apache937 i learned about that at installation n 2

[@swannschilling474]

Thank you so much!! Great find!!

[@WesRoth]

Happy to help!

[@killianlucas]

thanks wes, great video!!

[@WesRoth]

My pleasure!

[@muhammadazeemsadiq]

Excellent work

[@apache937]

This has plenty of potential. It's a shame that GPT-4 is so expensive, and NVIDIA is not producing high VRAM cards for consumer use.

[@WesRoth]

yeah, I spent $2 on OpenAI credits for that video (some of the stuff I tested was cut out). That doesn't sound like a lot, but it starts adding up really fast...

[@ITMTS]

Would be wonderful if you could showcase the local a bit more. Is it anywhere close to gpt4? And perform some commands like you would with openai api. Like, modify a document etc? Or summarize a PDF file. And would be wonderful to know if it also supports other languages other than English

[@wicked-admin]

Expensive how? $20 a month is not horrible..

[@apache937]

not chatgpt plus but the api if you want to use it plenty @@wicked-admin

[@AliAbassi1]

Thanks Wes 🔥

[@bzben9850]

Great Video!

[@WesRoth]

Glad you enjoyed it

[@kevinhaze796]

i spent a day...then i found this video.....thank god for you sir

[@tsadigov1]

You really saved my time thanks

[@WesRoth]

nice!

[@learning_AI]

Great video 👍

[@WesRoth]

Thanks 👍

[@personone6881]

Hey man! This is an awesome vid - Good Job!! - I was hoping to ask you a semi-related question if you don’t mind…? Since Code-llama is so large, and considering I’d want to probably install the 128B params HQ model to run locally, is it possible to install either a) Code-llama, b) Open Interpreter, or c) Both, onto an external drive? OR - Is that just stupid thinking cos it wont work through the cmd console anyway? - If for instance I can install Both (following option “C”) on an external drive and parent folder, hypothetically let’s say, “K:\InterpreterProject” I’d still just follow your instructions in this video just replacing your ‘C:\Users’ path with my ‘K:\InterpreterProject’ location, correct? I imagine it would be slower, but that is okay, something I have to accept given my lack of storage space remaining on my machines ‘C:\’ Would this work? How/where can I learn about what other possible problems I might encounter if I attempt to install it as described?

[@fuzzyorangetv]

thankyou i was having trouble with code llama

[@luckyluc9972]

Would this thing be able to access and use programming to interface with your computer files? I think it would be really cool to train a personal AI that organizes stuff on your PC

[@WesRoth]

yes. 100%. that's one of the big functions, you can say "take all my excel files in this folder and do XYZ to each" full guide coming soon!

[@joe_limon]

Can you revisit open interpreter if/when it can be run with a local llm such as code interpreter 34b

[@lrkx_]

This just made my day! No idea how you found this, but thanks for the heads up; it’s saved me ages in aider coding time trying to replicate CI!

[@WesRoth]

Great to hear! yeah, this thing is brand new and VERY exciting.

[@CM-zl2jw]

@@WesRothhow can I contact you ?

[@jersainpasaran1931]

can we use open interpreter, currently with chatgpt 3.5 turbo, how and where would the user be prompted?

[@cryptogrinch2978]

after i set the open at api keys . i some how deleted the keys in my openai dashboard. how do i add new keys and reset new keys?

[@AkaThePistachio]

I use WSL and would like to use it as part of my Linux subsystem within windows but when I run any command after using "interpreter" it throws back some error. ANy idea how I could fix?

[@MichelCDiz]

Does this local feature able to read local stuff? PDFs, Markdown, TXT?

[@RZRRR1337]

Is it free to use GPT-4 via the OpenAI API key? What does it cost?

[@guybarel9798]

tried to install code llama but i got the same problem the guy you showed had. (i have vs code, gcc, python, installed so idk why it doesnt work)

[@project-asgard]

very nice, but might get very costly if you use GPT-4. I bet if I use open-interpreter as much as I use the official one, it will cost me a fortune in API calls :D

[@apache937]

Steal someones api key

[@WesRoth]

yup, this vid cost me $2 in API calls. That really adds up fast, especially with something like this where it can just keep running and iterating.

[@user-js2gq2tr8n]

Why am I not prompted to enter the openai api key when I log in? I don’t know where to enter the api key, and if I ask a question directly, the result is a timeout.

[@brianmi40]

Hoping you or someone can help: I have Python 3.12 freshly installed via Windows (11) store and it has set an environment path to execute anywhere. I can readily run and exit Python with Ctl-Z.. When I do "pip install open-interpreter" the first line states "Defaulting to user installation because normal site-packages is not writeable" but runs to completion. After install completes with only the error: " Building wheel for tiktoken (pyproject.toml) did not run successfully." which seems to be common and non-critical, however, I can neither run "interpreter" afterwards ('interpreter' is not recognized as an internal or external command, operable program or batch file.) nor can even find it anywhere on my system as it appears it would be installed in C:\Program Files\Windows Apps which I cannot get access to as Admin. Your install guide mentions nothing that would seem to help. Unfortunately I'm neither a programmer (Some Fortran DECADES ago so not utterly clueless about it, but no modern languages) or experienced with Python environment, but am very Cmd prompt friendly and Windows knowledgeable in general.

[@buttpub]

open interpreter is the next level, but as long as its bound to OpenAI's api its unusable, it needs to work with llama2 and falcon as well

[@angloland4539]

❤

[@alexandrupop7461]

hey guys! do you have any idea why is not working for me after having everything installed? 'interpreter' is not recognized as an internal or external command, operable program or batch file.

[@lomek4559]

onnxruntime package requires numpy>=1.24.2 while tensorflow-intel 2.12.0 package requires numpy=1.22, I'm unable to install it even on anaconda. What to do?

[@wvagner284]

Nice video! Tks! Is it possible to run it with a web ui to be able to upload files?

[@WesRoth]

it has access to all the files on your computer if you're running it locally. You can also use google collab and looks like they just launched Azure support this morning. But, nothing like a web ui yet (unless I'm missing something)

[@mohammadfawaier6080]

I followed all the steps, but I got this error message: 'interpreter' is not recognized as an internal or external command, operable program or batch file. Any possible solutions??

[@DignitNetwork]

can u tell me how to build the gameboy on the controller

[@MrGnolem]

How does it handle big repositories? Thats where the crown jewels are atm. Starting small: build a snake game, fetch me a joke, etc. I believe. But now for professional usecases making our life easier 😅

[@EricPena86]

How do you upload files like pdf or csv in Terminal for gpt-4 to analyze?

[@WesRoth]

it has access to your file, no need to upload, just ask it. for example, if you have an csv file on your desktop, ask it to list all files on your desktop, then ask it to analyze the csv. It will build the code it needs to access it and analyze it. PS if you meant in google collab, then do the following: Open Google Colab and click on the "Files" tab in the left-hand side of the screen. Click on the "Upload" button in the top-right corner of the file browser.

[@EricPena86]

Awesome! Thank you so much!

[@normanlove222]

I could not get code llama to work on Windows 10 following your example. says llama not installed properly etc. Maybe you should share your setup, GPU, RAM etc

[@WesRoth]

Check the link I have to GitHub with the "issues": github.com/KillianLucas/open-interpreter/issues/63 This one has the solution I used and also some other people listed theirs. The developer is actively working on it, so might be a fix soon. Here's what another user on there said: __________________ in Windows 11 environment: install Visual Studio 2019/2022 start a Developer Command Prompt (optional) conda activate env pip install llama-cpp-python works like a charm ___________ let me know if that works!

[@mrlsessions6807]

'interpreter' is not recognized as an internal or external command, operable program or batch file. how do i fix this

[@apache937]

pip install

[@WesRoth]

I had the same issue at first. not sure what fixed it, but here's what I tried: -update Python -re install Python -reboot comp Type in the exact message into gpt4 and it will give you some things to try, I think this had to do with the Python environment. I think I also had to reinstall python at some point because pip stopped working, I used this: python -m ensurepip --upgrade I will be posting a more in depth intallation tutorial with common troubleshooting issues soon, that will be available here: natural20.beehiiv.com/ (give me a day or two)

[@milosjovanovic803]

I just tried to install open Interpreter and instead of just install it, my computer was start uninstalling everything including openAI. Why?

[@mbrochh82]

so, does this actually work with a local model? the teaser video is so over the top, there is no way that it actually works as shown in the video.

[@WesRoth]

I have my doubts about Code Lama being on the same level as GPT-4. I will have to check, but I would be shocked if it's anywhere close. (but, better open source models are coming soon)

[@tubesarkilar]

nice

[@EnjoyCocaColaLight]

Why do you always use cmd prompt to install things?

[@WesRoth]

There are other ways, but this is the easiest one. The developer did a great job making the installation super simple.

[@jorgeromero4680]

i tried the open interpreter local (34B) and found something very weird. On GPU (Nvidia 4080) runs extremely slow consuming all resources GPU memory, cores at 100% + consumed shared memory 32GB + CPU! When i run it on CPU runs faster than the GPU with less resources. Very weird.

[@ITMTS]

Can you also share a bit about your experience with locally?

[@jorgeromero4680]

@@ITMTS what you need to know?

[@yushengyang2]

What LLM are you using if It is not OpenAI chatgpt!

[@user-mv9ul9tz1c]

Sorry, I'd like to inquire about installing Open Interpreter. I haven't entered an API Key and I'm directly using Code Llama with 7B data, medium mode, and CPU (not GPU), and I have downloaded the models. However, I'm encountering an error during installation with Code-LLama. The error message is: 'Command '['C:\Users\XXXX\AppData\Local\Programs\Python\Python310\python.exe', '-m', 'pip', 'install', 'llama-cpp-python']' returned non-zero exit status 1. Failed to install Code-Llama. We have likely not built the proper Code-Llama support for your system.' But when I run 'pip install llama-cpp-python', it shows below error note: This error originates from a subprocess, and is likely not a problem with pip. ERROR: Failed building wheel for llama-cpp-python Failed to build llama-cpp-python ERROR: Could not build wheels for llama-cpp-python, which is required to install pyproject.toml-based projects Thank you

[@WesRoth]

Here's the thread for that issue: github.com/KillianLucas/open-interpreter/issues/63 People have suggested a few solutions. If those don't work, please post exactly what you did in the comment, to that specific thread. This might be an issues that's resolved within a few days. Sounds like the dev is mainly on Mac and trying to fix this issue for other OS.

[@user-mv9ul9tz1c]

@@WesRoth Thank you, It seems that currently, this resource is primarily suitable for use on MAC computers. However, you can certainly try using a GPT-4 key for different modes. Using an API key that requires payment may result in fewer issues.

[@zakaria20062]

I think we need some video non-openAI

[@vicarioustube]

Can I ask a stupid layman question. What can this do for a 3D designer like myself who isn't into coding. I clicked because I thought it was going to INTERPRET something of my choice. Can it do anything locally with graphics.

[@AlienAnthony]

Ita a cool idea but code lama isnt the best at it. I hope it gets better. Gotta remember if you do use code llama the token limit is very small. Ittle crash after about 4 commands.

[@WesRoth]

I'm most excited about Open Interpreter plus GPT-4. I'm hoping code lama does well, but I'm not betting on it.

[@googleSux]

So the truth is its not that easy to get to install or run under windows and LLama. With the API its not a problem but you can easily spend 10 bucks a day on that. The LLama version was pretty tricky, first the LLama-cpp-python did not install for the longest time (updating PIP, GIT) then would not compile, hat to install a specific version of VS). Nothin for amateuers. The free version now works but sooooo sloooooow something cant be right still.

[@rubeng4833]

guys BE CAREFUL with openAI, these boys dont care about your usage limits. you can blast way past you limits because there's a so called bs 5 minute delay. and guess what? support is AI hahaha

[@williamstamp5288]

Yooo

[@WesRoth]

ikr

[@williamstamp5288]

I created an open ai chat assistant that uses STT and TTS. This will give me that much more control over my dev process, especially when I tailor it to me and give it persistant memory. Let's goo

[@dawid_dahl]

Woah. This thing literally has a greater-than-zero chance of destroying your entire computer and contact any malicious external website or system every time you make a prompt. That said, amazing tech! But I think I’ll stick to OpenAI’s sandboxed terminal for now. 😄

[@yagoa]

not free with cgpt

[@WesRoth]

it's free with code lama or you can pay for gpt4. the new thing is that you can run it free with an open source model

[@yagoa]

sry I just saw that you were taking the cgpt route @@WesRoth