How to setup Auto-GPT on macOS 💽

Read on to find out how to easily install Auto-GPT on your Mac using Python, plus additional tips & tricks.

---

After a month of scrolling past the occasional tweet that referenced Auto-GPT and all of the cool things it can do, I finally decided to give it a go.

I've simplified and tailored the instructions specifically for macOS, and have done my best to make it accessible for anyone non-technical.

Link to this headingSteps to set up Auto-GPT on macOS:

1. Get the code
2. Set up your OpenAI account
3. Setup your Auto-GPT environment
4. Install Python 3
5. Run Auto-GPT
6. Bonus tips & tricks

Link to this heading1. Get the code

Download Auto-GPT from this zip file or git clone the [Auto-GPT repository}(https://github.com/Significant-Gravitas/Auto-GPT).

Link to this heading2. Set up your OpenAI account

1. Sign-up for a developer account at OpenAI(https://platform.openai.com/overview) if you don't already have one.
2. Create an OpenAI API key if you don't already have one.
3. Make sure you have a paid OpenAI account by going to Billing within your OpenAI account.

Link to this heading3. Setup your Auto-GPT environment

If you unzip the zip file you downloaded/cloned in step 1, you should see a folder called Auto-GPT-master.

Within that, you should see a file called .env.template. If you don't see it, you might need to enable viewing hidden files on your Mac.

Now you can:

1. Rename .env.template to .env
2. Open .env in a text editor
3. Update OPENAI_API_KEY=your-openai-api-key to be the API key you setup in the previous step (no quotes or anything after the =): OPENAI_API_KEY=skfeafaefefefesfsf

Link to this heading4. Install Python 3

You likely already have Python 3 on your Mac, as macOS installations have been including Python 3 since 2018.

If you're not sure, you can check by running python3 --version in your terminal. If you see something like Python 3.10.10, you're good to go.

Otherwise, you can install Python 3 with homebrew.

Link to this heading5. Run Auto-GPT

1. Open your terminal
2. Navigate to the Auto-GPT-master folder you downloaded/cloned in step 1 (e.g. cd ~/Downloads/Auto-GPT-master in terminal)
3. Run ./run.sh in terminal

• Note: if you don't have access to GPT 4, you should instead run ./run.sh --gpt3only in terminal. If you're not sure, use the gpt3only flag anyway.

You should see something like this:

Missing packages:
beautifulsoup4, colorama, distro, openai, playsound, python-dotenv, pyyaml, readability-lxml, requests, tiktoken, gTTS, docker, duckduckgo-search, google-api-python-client, pinecone-client, redis, orjson, Pillow, selenium, webdriver-manager, jsonschema, tweepy, click, charset-normalizer, spacy, en-core-web-sm, coverage, flake8, numpy, pre-commit, black, isort, gitpython, auto-gpt-plugin-template, mkdocs, pymdown-extensions, openapi-python-client, pytest, asynctest, pytest-asyncio, pytest-benchmark, pytest-cov, pytest-integration, pytest-mock, vcrpy, pytest-recording
Installing missing packages...

And then after a few minutes, you should see something like this:

Welcome to Auto-GPT!  run with '--help' for more information.
Create an AI-Assistant:  input '--manual' to enter manual mode.
  Asking user via keyboard...
I want Auto-GPT to:

Now that you have Auto-GPT running, try giving it a few fun easy tasks to start, eg I want Auto-GPT to: summarize the latest news in the United States or `

Link to this headingPotential issues

1. python: command not found

If instead you see

./run.sh: line 2: python: command not found
./run.sh: line 8: python: command not found

Then the simplest solution is to open run.sh in your text editor, and replace each instance of python with python3 and pip with pip3 and then re-run ./run.sh.

2. DOUBLE CHECK CONFIGURATION error

This can happen for at least two reasons: either you don't have a paid OpenAI account, and/or you don't have access to GPT 4.

Make sure you have a paid account with OpenAI. When I first started trying out Auto-GPT, I kept getting an error message that said:

DOUBLE CHECK CONFIGURATION  Please double check that you have setup a PAID OpenAI API Account. You can read more here: https://docs.agpt.co/setup/#getting-an-api-key

It took me a minute to realize that the Auto-GPT docs also reiterated this point and that I did not in fact have a paid account with proper billing info attached 🤦‍♂️

If you don't have access to GPT 4, then use ./run.sh --gpt3only instead of ./run.sh to run Auto-GPT.

Link to this heading6. Tips & tricks

1. Be specific. If you want Auto-GPT to use the internet or to write to a file, it can help to specify that.
2. Think big with your tasks. You can use Auto-GPT to help you:
   • buy products online
   • to do research synthesizing a bunch of search engine results into a single document since it can make Google requests itself
   • to write a blog post, a song, or code.
3. Checkout the Ideas section of the Auto-GPT discussions for inspiration: https://github.com/Significant-Gravitas/Auto-GPT/discussions/categories/ideas

Link to this headingAdditional resources:

• Auto-GPT docs: https://docs.agpt.co/
  • the docs are useful for learning how to use text to speech with auto gpt, to learn more about how it works, or for additional use cases
• OpenAI API documentation if you want to learn more about how that works: https://beta.openai.com/docs/api-reference/introduction