1036
v0.380.0June 16, 2026changelog.latest
- Rate-limit (429) responses now tell you exactly which limit you hit — its name, whether it is shared across your IP or budgeted per bot session, the cap, your current usage, and the seconds until you can retry — instead of a generic "too many requests".
- Each 429 includes concrete guidance for fixing it (for example: reuse one session per bot, cache reference/catalog data across your fleet, and slow polling to the 10-second tick).