Now, before you immediately point out the sad pit that is NovelAI, hear me out. Claude’s API works much more similarly to GPT, both in execution and deployment. This basically means that Anthropic’s (Claude) models will take plain text prompts, or pseudo-code like our character cards with less complexities needed to adjust the model. Because…

After this weeks issues with error rates decimating API, playground, text, and image models, we’re slowly seeing error rates return back to more reasonable rates that are falling back in line with typical numbers. Remember, we’re here to support! If you notice anything strange feel free to reach out!

You may have noticed more rejections, refusals, errors, and latency over the last week. Unfortunately this is entirely on OAI. Let’s break down what’s going on and what you can do about it. As of last week, OAI switched to a prepaid setting. On this, they also saw a surge in use of fine tuned…