Token limit rate problem

[therenansimoes]

I'd like to suggest an option on menu to set a max tokens per minute on requests to don't loose the conversation because of rate limits. Or, if codex get rate limit error, just wait instead of crashing and loosing session

[jrjones]

Even simpler, throw a --ratelimit argument that specifies a usage rate tier, and can have simple internal logic throttle accordingly. https://platform.openai.com/settings/organization/limits

Or could even extract the data from the HTTP headers if I'm not mistaken. https://platform.openai.com/docs/guides/rate-limits#usage-tiers

I won't have time to look into this until next weekend, but seems achievable.

[CosmicCompassAO]

Ya this needs to be fixed. When working with a big file you literally just waste money.