Warning: The file 'auto-gpt.json' does not exist. Local memory would not be saved to a file.

[draggingsurf]

Duplicates

• [X] I have searched the existing issues

Steps to reproduce 🕹

Warning: The file 'auto-gpt.json' does not exist. Local memory would not be saved to a file.

Current behavior 😯

Expected behavior 🤔

No response

Your prompt 📝

# Paste your prompt here

[lalaland231]

hey bro,you should install the 'tweepy' on python

I did install it not working.

[lalaland231]

Python -m AutoGPT

我做了一切，但它仍然给了我这个。我也尝试了python -m autogpt --debug并且不起作用。

It looks like you're missing some dependencies. You should run this command in the root directory: pip install -r requirements.txt

I run the pip install -r requirements.txt, and it installed somthing, but is still mot working.

[unclezhang0309]

Python -m AutoGPT

我做了一切，但它仍然给了我这个。我也尝试了python -m autogpt --debug并且不起作用。

It looks like you're missing some dependencies. You should run this command in the root directory: pip install -r requirements.txt

I run the pip install -r requirements.txt, and it installed somthing, but is still mot working.

---

It seems like there's an error with your Git executable. I'm not sure how you've been operating it before, but you can try troubleshooting the issue by referencing the error messages and prompts.

[mindfu23]

OK, whew. Just solved this for myself also. As per above, just to sum up:

1. Create an empty file in the main directory, auto-gpt.json
2. change the .env file. Remove the .template extension. And make sure it's associated with text file or similar.

Note that Mac, for whatever reason, can keep adding the .template extension after you remove it. You have to command i, look at the properties, remove the extension there and confirm you want to remove the extension.

Woo. Thanks all :) I now appear to have this working on my Macbook. M2.

[noelabey]

OK, whew. Just solved this for myself also. As per above, just to sum up:

1. Create an empty file in the main directory, auto-gpt.json
2. change the .env file. Remove the .template extension. And make sure it's associated with text file or similar.

Note that Mac, for whatever reason, can keep adding the .template extension after you remove it. You have to command i, look at the properties, remove the extension there and confirm you want to remove the extension.

Woo. Thanks all :) I now appear to have this working on my Macbook. M2.

I tried this and the issue was solved for me too. I ran it on an M1 Macbook with python3 -m autogpt

[chunqiushi]

give up

[amandoabreu]

I get this when I run it in a docker container

Did you figure it out? Most here seem to not be running inside docker. I notice that if I access the docker container shell and run python -m autogpt from within, it works.

However to copy the files from the docker container into the host I need to use: sudo docker cp <container id>:/home/appuser/auto_gpt_workspace/ .

Run docker ps to get container ID

[gavinuhma]

If you are having this issue with docker compose:

From the main project directory run touch auto-gpt.json to create the empty file.

Then open Dockerfile and add the following line which will copy the file into the docker container during build:

COPY --chown=appuser:appuser auto-gpt.json .

Then build:

docker-compose build auto-gpt

[MCplayerFromPRC]

If you are having this issue with docker compose:

From the main project directory run touch auto-gpt.json to create the empty file.

Then open Dockerfile and add the following line which will copy the file into the docker container during build:

COPY --chown=appuser:appuser auto-gpt.json .

Then build:

docker-compose build auto-gpt

It works, thank you. But do I have to use docker cp command to copy it out of the container before I cut it off. @gavinuhma

[hatgit]

~Does anyone know whether we can we specify in this json file where the stable branch should save files, since there is no workspace directory like there is in master branch? Or know how to otherwise save/find files when running stable version in docker environment?~ I ended up just creating a workspace folder and that resolved it.

[hatgit]

P.s. When trying to run workspace.py in stable which should create workspace path, getting error:

File "workspace.py", line 18
    def path_in_workspace(relative_path: str | Path) -> Path:

[tchwpkgorg]

I've created an empty "auto-gpt.json" file in project root, but still getting "Warning: The file 'auto-gpt.json' does not exist.":

$ ls -l auto-gpt.json 
-rw-r--r-- 1 chatgpt chatgpt 0 Apr 27 08:57 auto-gpt.json

$ docker-compose run --rm auto-gpt
Creating auto-gpt_auto-gpt_run ... done
Warning: The file 'auto-gpt.json' does not exist. Local memory would not be saved to a file.
(...)

Edit: missed this comment - https://github.com/Significant-Gravitas/Auto-GPT/issues/1410#issuecomment-1522751914 - which fixes the issue.

[LeonKnook]

I don't like it at all, sorry. I got everything working. I also paid. Keys for autogpt, elevenlabs, google all fine implemented and without remarks for them. Then put everything to work. First with 5 questions in continuous mode. AutoGPT is working for hours without writing anything to files, doing the same attempts over and over again with google searches and what not. Then I thought, let me just ask 1 question then it might take less time. Same story, trying again and again on google. You can see in the scrolling text that he keeps visiting other websites, but there is no end to this. I'm in this for days and at first it seemed insane, until eventually you don't get the results you want and slowly get frustrated. I take a break and go to other AI applications that do work and are a lot more user-friendly. Yes, the old dosbox work slaves can really enjoy themselves here ;-)

[hatgit]

@LeonKnook Just FYI, the continuous mode is not advisable given how buggy the app can still be, I find automating 5-10 commands at most if the goals are clearly defined, can help before an error is encountered, and if you notice it trying to do something and getting an error, either have it avoid doing that behavior again by explicitly tell it not to, or try to figure what is causing error or change the goal so that it avoids that behavior.

[FamSCR]

Found this on a website (https://prinspeter.medium.com/how-to-use-autogpt-using-docker-a-step-by-step-guide-34ad33957de9) and once updated it worked.

[pabloflego]

Found this on a website (https://prinspeter.medium.com/how-to-use-autogpt-using-docker-a-step-by-step-guide-34ad33957de9) and once updated it worked.

That addressed the issue of the LocalStorage Database not present and effectively keeps the data saved outside the container. Additionally I've updated the volumes to look like this:

    volumes:
      - "./autogpt:/home/appuser/autogpt"
      - ".env:/app/.env"
      - "./auto-gpt.json:/home/appuser/auto-gpt.json"

This is because docker-compose is mounting the app in the root /app but the app seems to being run from a stale copy from /home/appuser/autogpt which I don't understand why. But for most of the cases just mounting the json file is enough.

[github-actions[bot]]

This issue has automatically been marked as stale because it has not had any activity in the last 50 days. You can unstale it by commenting or removing the label. Otherwise, this issue will be closed in 10 days.

[github-actions[bot]]

This issue was closed automatically because it has been stale for 10 days with no activity.