I am looking for a chatbot, where I can insert my OpenAI API key and not face any limits in terms of input and ouput.
You need to define what you mean by “not face any limits”.
If you are using OpenAI services, you will not be able to generate policy-violating content without threat of termination.
If you are using OpenAI language models, they have been trained to deny objectionable content.
If you are using the OpenAI API, there are rate limits and spending limits per account.
By no limits, I mean the ability to make a 20,000-word prompt and receive a 20,000-word reply.
All AI models have a limit called context length. This is the computationally-expensive memory where the AI will both receive its input and produce its outputs.
The only AI that is anywhere close to the 40000 words = 60000 internal AI tokens is Anthropic’s Claude-100k. And it is limited use by approval for high-profile customers.
I can make it work with 32k tokens via GPT4. However, is there a tool where I can connect my API key and work in a chatbot interface very similar to chatGPT?
Currently, the only methods for accessing GPT-4-32K are applying for an Enterprise ChatGPT account Introducing ChatGPT Enterprise or applying for and being granted Azure OpenAI Services GPT-4-32K access.
Is this what you mean?
Of course you still need the gpt-4-32k API which is readily available on Azure these days.