Promptwatch Logo
All updates
Bug Fix

Bug Fixes & Stability Improvements

A round of fixes across the platform to keep things running smoothly:

  • Monitor comparisons measure the selected range against a previous period of the same length. Last 14 days used to be compared with the 15 days before it, which skewed changes in visibility, share of voice, and position.
  • A prompt's average visibility in the prompt list fills in after its answers are scored again, instead of staying blank until the prompt's next run.
  • Exported citation positions match the Citations table. They were one too high before.
  • Tracked Claude, Claude Code, and OpenCode answers no longer repeat our search provider's quota errors. When web search is down, they say so in a neutral message.
  • Perplexity tracking picks up answers again after Perplexity changed the address of its answer pages.
  • API keys and MCP tokens follow the same access rules as the dashboard. They stop working once a canceled subscription ends, and work again when you resubscribe.
  • To keep the API fast for everyone, responses over 100 KB are no longer cached, Business allows 250 requests a minute, and Enterprise allows 400 a minute and 24,000 an hour. A raised hourly quota now raises the per-minute limit with it.
  • The marketing site no longer crashes when Google Translate is on.

Start Tracking Your AI Visibility

Join thousands of companies using Promptwatch to monitor and improve their presence in AI search results.

Promptwatch Dashboard