What each engine probes

Five LLM engines, when each one is enabled, and how to choose which to run.

SyncSEO probes five engines, all reachable from Settings → Engines:

  • ChatGPT — GPT-4o family. Default enabled on every plan.
  • Claude — Anthropic Claude Sonnet (and Opus on Premium-tier article generation). Default enabled on every plan.
  • Perplexity — Sonar models, citation-grounded. Default enabled on every plan.
  • Google Gemini — Growth plan and above.
  • Google AI Overview — Growth plan and above.

Toggling engines

From Settings → Engines, check or uncheck per workspace. Empty selection means “use the system default” (the three free-plan engines). If your buyers don’t use Perplexity, turn it off and save credits — runs scale with engine count.

How probes are sent

Every engine receives the same probe variant for a given prompt (D1 cold, D2 regional, D3 niche-contextualised, D4 alternatives). We don’t hand-tune per engine; the same question goes to every engine and the comparison is apples-to-apples.

Why not all five always?

Cost. Each probe call costs a small amount through OpenRouter; a 200-prompt run × 5 engines = 1000 probes. We bake the cost into the credit price for plans that include all 5; Free / Starter stays at 3 to keep the per-credit math reasonable.

← Back to knowledge base

Question we didn’t answer? Email hi@syncseo.io.