abielzulio
commented
1 year ago Hey @abielzulio
Love the extension.
After a longer conversation I ran into an error, caused by hitting the token limit. This pull request attempts to fix this error.
A pretty naive implementation, but I think it's better than having to stop the conversation. It approximates the number of tokens by counting the words, and then removes old messages to stay under the limit (3750 tokens).
I've also added some other feats:
Fetch models from the API. This allows to use GPT-4 if enabled on the user's account.
Show the error message coming from OpenAI instead of the generic error message.
Let me know what you think.
Hey! Thanks for your PR!
We’re actually one step closer to getting it published to the store (you can check my PR here: https://github.com/raycast/extensions/pull/5253) and we already had some changes.
Could you open a PR from my raycast/extensions fork at https://github.com/abielzulio/extensions (branch: chatgpt/init)? So I can merge it into my existing PR
CruelMoney
Hey @abielzulio
Love the extension.
After a longer conversation I ran into an error, caused by hitting the token limit. This pull request attempts to fix this error.
A pretty naive implementation, but I think it's better than having to stop the conversation. It approximates the number of tokens by counting the words, and then removes old messages to stay under the limit (3750 tokens).
I've also added some other feats:
Let me know what you think.