Repository navigation
How to implement continuous dialogue with api #164
Description
Activity
Hi @winddack! Thanks for writing in.
Unfortunately, there is no conversation_id parameter for our API. Our official API documentation can be found here: https://beta.openai.com/docs/api-reference/completions.
You should be able to able to achieve what you're looking for by sending in multiple requests, each with the previous prompt+generation as the new prompt.
You can also find more help in our discord community, help.openai.com, or community.openai.com
Reacted by HenryReacted by Henrythank you,i will try it
@hallacy How can we hold longer conversations this way? If we have to provide the full history every time, we will very quickly hit the 4096 prompt token limit.
@hallacy thanks for the answer, just to confirm: the
idwe get from aopenai.ChatCompletion.createresponse has nothing to do and cannot be used asconversation_id?You should be able to able to achieve what you're looking for by sending in multiple requests, each with the previous prompt+generation as the new prompt.
While that definitely works (thanks), it is very expensive because each time the entire previous conversation is counted as new input tokens isn't it? Would be great to have a more efficient way to have a continuous conversation/dialogue.
- added a commit that references this issue
on Apr 27, 2025
How to use the api to achieve contextual continuous dialogue like ChatGPT , such as the first call to the API to ask you what your name is, and the second re-request to get a different result. I tried to use ChatGPT to answer this question and it told me that I could take a conversation_id parameter, but I was prompted with an error when I actually used it.please help me, This is important