I have a dev tier account, and when I use tools like AI-FDE, I feel the rate limits are low enough to prohibit productive development.
I appreciate this is a free service, so I want to understand if there’s a way for me to use my own Claude API credentials for AIP? That way I get billed for it directly. Or if there’s a way for me to set limits and just pay per usage, stilling keeping the other platform limitations (user count etc.) such that I won’t actually go to production with it.
Whilst BYOM is technically possible with the four primary model families, I’m pleased to report that we’ve increased the rate limit thresholds for AI FDE compatible models on Developer Tier accounts. Please go ahead and retry AI FDE and hopefully you’ll find it a much better experience.
For full details on the individual limits per model, checkout “AIP Usage and Limits” in the Resource Management Application.
I’m doing a fair bit of work in AI FDE with Sonnet 5 in my stack right now, I have tweaked the settings in “manage rate limits” and “reserved capacity” which has helped a bunch with rate limit hits, but I also see a discrepancy in what’s shown in Resource Management
What is the actual limit on my stack? 1m tpm or 400k tpm? I think it’s more like 400k, but would be good to have that confirmed
