Receiving of of Code Interpreter

Dear Developer Community,

I am Alexander, an active participant in Mimo Incorporated. We are engaged in various operations, one of which is a considerable involvement in scholarly research relating to Artificial Intelligence, especially concerning educational methodologies.

In line with our ongoing work, we are looking forward to integrating the Code Interpreter into our research process. We believe it will significantly aid in the progress of our research, bringing us closer to achieving our desired goals.

Hence, I am writing to inquire about the anticipated timeline for the release or availability of the Code Interpreter. Understanding this will assist us in planning our research progression and ensuring that we are well prepared to implement it once available.

Thank you for your time and consideration. I look forward to hearing from you soon.

Best regards,



Goodluck :slight_smile: I am also patiently waiting.


you too, also non business related, have you received browsing, and/or plugins

1 Like

I am a Plus/Pro user. I think all have access to it now, you have to enable it in settings. :slight_smile:

1 Like

I have it, I was just seeing that some people still don’t have access :slight_smile:

We are all waiting for Code Interpreter… :sweat_smile:


As I wait for access, I have also started actually doing it, but with PaLM 2. I have to ask you folks since you seem really into Code Interpreter …

Since it’s apparent that OpenAI must provide an entire compute stack for Code Interpreter to work, how does that compare to an AI company that already has the compute stack and 30+ years experience running code?

It seems to me that this is an uphill battle for OpenAI, but a bit of a glide for Google.

Bear in mind - I am a complete idiot with complex ideas like this, but I was able to make something work. Here’s how…

In everyday Google Apps Script code, I have built a very rudimentary “code interpreter”:

  • PaLM 2 takes in code-building instructions (in a simple chat UI)
  • It generates the code
  • Hands it off to my code “runner” which executes it
  • Returns with results, repairs it, rinse-repeat

Aside from the compute stack provisioning, how is this not the same? What am I missing?

As a follow-on question concerning AutoGPT -

  • I created a simple spreadsheet listing all of the tasks I want to perform
  • Each task includes a prompt and possible branching flows based on results
  • I built a simple automation harness to fire off each of the prompts in sequence or branch/goto as needed
  • Results are generated in a Google Doc
  • This AI automation process is wired to a script trigger - it runs every night to create an automated briefing about our competitors

I’m an AI dummy, but I was able to do this in a day. Certainly, it’s a one-off, however, it’s clearly an approach to AI automation that seems to have legs.

How is this any different from AutGPT?


I’m a Plus user and I don’t have it yet, just looked at my settings.

Code Interpreter is by far the most exciting thing yet. Patiently waiting :weary:


Hi, I have access to the other plugins & web search but not for the code interpreter, could be a bug? I wanna use it in my GPT plus suscription.
It will be a pleasure if you please enable it in my account
Thank you for your time

Sometime people don’t read and keep asking the same question again.
The Code Interpreter as per their announcement is still in AlphaPhase and they asked folks to be patient, please search the forum and you would see OpenAI staff response to that.

The other piece is people asking to expedite for research purposes, for that I’d say one thing.

You can build a much more powerful code interpreter if you have Dev plugin access.
It will only take few amount of code and you can ask GPT-4 to do that for you too if you don’t want to write it yourself :slight_smile:
You would get far better integration since you have control over the environment with all libraries you want and more importantly, enough firepower to run some real stuff on your machine without relying on the sandbox environment.

I might create a demo for this, but it’s really not a rocket science to build a universal interpreter with what GPT-4 and existing plugins offer at the moment.

Those are my 2 cents on this subject.



How did you connect AutoGPT to google doc? I usually save it in the autogpt directory called “work space” or something like that

I didn’t - I created my own (very baby-like AutoGPT influenced by @stevenic and @curt.kennedy) in Google Apps Script. My approach is designed to publish results to Google Docs and Google Apps Script has direct SDK access. Google I/O demonstrated many of the advantages of using adjacent Workspaces apps to create more business-friendly solutions. I’m following the easy path to connect smart AGI apps to existing business tools.


I’m still waiting too. As a Director of Analytics I’m really interested to see how much automation this can bring. Code Interpreter looks like it will speed up the basic work pretty quickly, leaving more time for complex/strategic analysis.