AutoGPT: Permanently Stuck "Thinking..."
⚠️ Search for existing issues first ⚠️
- I have searched the existing issues, and there is no existing issue for my problem
GPT-3 or GPT-4
- I am using Auto-GPT with GPT-3 (GPT-3.5)
Steps to reproduce 🕹
I have it installed and when I enter any prompt, it just says “thinking…” and never stops loading.
Current behavior 😯
It never stops generating a response and never actually produces anything.
Expected behavior 🤔
No response
Your prompt 📝
# Paste your prompt here
About this issue
- Original URL
- State: closed
- Created a year ago
- Reactions: 5
- Comments: 34 (8 by maintainers)
That’s because you don’t have a paid plan for the OpenAI API, regardless of whether you have ChatGPT Plus or not.
https://github.com/Significant-Gravitas/Auto-GPT#openai-api-keys-configuration
Ah, I assumed Pro implied “paid plan”. Thanks for pointing this out 😃
EDIT: Can now confirm, I setup a paid plan, generated a new API key and the stuck “Thinking…” problem went away.
If you want to use gpt to it’s full extend, they charge for that. That’s how businesses work. Them offering free trials is already pretty amazing. If you want to use auto-gpt you can also use a local llm model, the setup is harder though.
But do not confuse open source projects with being a company, we all contribute for free here.
@Pwuts thanks for the reply. I saw the hint ultimately, but since I already had the API keys and (seemingly) everything else I needed to get started, I didn’t bother to check the installation explanations until after I spent some time trying to get it to run. I just think that a little hint inside ‘requirements’ or ‘quickstart’ of Readme.md would help avoid this in the future. Still, it’s very easy to set up and very interesting, so thanks for the work you guys put in, I don’t want to sound ungrateful 😉
Many programmers are working on getting these bugs/features implemented. I have noticed this occurring, especially with selenium. There will be corrections in the future.
same here. after a long time it came back with a “RuntimeError: Failed to get response after 10 retries”. I have chatgpt plus.
same is happening to me- i am trying to fitget with it right now
#TWO COMPANIES IN PARTNERSHIP TO MAKE MORE MONEY, THAT’S IT!
That’s a shame… Might be a good idea to add this detail in the project requirements, no? Also, why is Auto-GPT dependent on a paid API even when --gpt3only is used?