Some runs were failing outright, or worse, quietly saving empty results, whenever one AI engine had a hiccup. Both are fixed, and engine selection is now yours to control.
What changed
- Each engine gets its own 45-second timeout, so one slow or unresponsive provider can't hold up — or hang — the whole run.
- If an engine fails, the run still saves results from every engine that succeeded, instead of discarding everything or saving a blank result for the one that failed.
- If every enabled engine fails, you get a clear error instead of a silently empty row.
- After a partial failure, a warning lists exactly which engine(s) didn't come back, so a missing result is never mistaken for "no mentions."
- Which engines you track is now a per-account choice in Settings, not a fixed set — pick the ones that matter for your brands.
Why it matters
AI providers have outages and slow moments like any other API. You shouldn't lose an entire run — or worse, get a misleadingly clean "zero mentions" result — because of one provider's bad day.