AutoGPT and GPT4Free

[StarGgazeER]

how can we generate an API key token , from GPT4Free which is going to run locally indeed of connecting with GPT4 Which for testing purposes which will be expensive , in such a way we will be able to have an open GPT4Free and AutoGPT , any suggestions ?

[fabriquepotentiels]

should be an update for sure ! as many can't afford to pay to do great stuff for the ecosytem !

[OSZII]

You can't! That's not how it works. AutoGPT calls the openAI api. Only openai can generate valid keys! However what you could do:
You could rewrite the part in autogpt that sends a request to openai to send a request to gpt4free. You would also need to figure out how to call gpt4free. But should easily be possible if there is no lawsuit against gpt4free

[xtekky]

working on it ; )

[nick-the-dev]

any news? :)

[proitservices]

One thing to cut cost I've put together is the embedding API replacement
https://github.com/proitservices/elmo_embedding_api
or:
docker pull peterromanowski/elmo-embedding-api:0.1.4

Any suggestions what to throw into it - just let me know

[juancruz1962]

We can use the same option used by chatgpt-web and chatgpt-next web using one-api using custom apis and endpoints similar to free-ai-apis. It's some sort of a single API for all LLMs, the same as what gpt4free is doing. One-api also supports some available apis from gpt4free providers. I've been using it on other UIs locally, and works fine. Using RapidApi can also be possible together with free-one-api, OneAPI GenossGPT and lots more. I only test GPT UIs on github combining features that adapt to its settings starting from xtekky's project. I hope this perks an idea. Thanks.

[Woukim]

I'm trying it and it doesn't seem to be working right now.