Supported AI platforms
Which AI platforms Obsurfable monitors today—ChatGPT and Gemini—how per-platform scoring works, and what is coming next.
Overview
Obsurfable runs your monitored prompts against real AI platforms and stores each answer separately, so you can see how the same buyer question is answered differently depending on where it is asked.
| Platform | Model | Status | Web search |
|---|---|---|---|
| ChatGPT (OpenAI) | GPT-5.4 mini | Live | Yes (Pro) |
| Gemini (Google) | Gemini 3.6 Flash | Live | Yes, via Google Search grounding (Pro) |
| Claude (Anthropic) | — | Coming soon | — |
| Perplexity | — | Coming soon | — |
Platforms that are not yet wired to a provider appear as Coming soon badges on the AI Brand Health dashboard.
Why cross-platform monitoring matters
Each AI platform reads a different search index and applies different citation preferences. ChatGPT retrieves through Bing and OpenAI's own crawl; Gemini grounds on Google Search. The same prompt can mention your brand on one platform and omit it entirely on the other, so a single-platform score hides half the picture.
Choosing a platform for a run
On a prompt run, the Ask button has a platform picker next to it. Pick ChatGPT or Gemini, then ask an individual prompt or run them in bulk. Results are stored per prompt per platform, so asking the same question on both platforms gives you two independent answers to compare rather than overwriting the first.
Each result card shows a platform badge when a run contains answers from more than one platform.
Scheduled runs
Scheduled automations run every prompt on every platform enabled for your company. Choose them under Monitored AI platforms in Settings; at least one platform is always required.
Companies default to ChatGPT only, so adding Gemini to scheduled runs is opt-in — each extra platform runs every prompt again and multiplies the work a scheduled run does. This setting does not affect asking prompts by hand: the picker on a prompt run always offers every available platform.
Per-platform scoring
Obsurfable stores a separate AI Visibility Score series for each platform plus a blended overall score:
- The overall score is what you see on the dashboard gauge and trend chart by default.
- Per-platform scores appear above the main report when you have results on more than one platform, each with its own delta.
- Incidents are raised per platform, so a visibility drop on Gemini does not get averaged away by a stable ChatGPT score.
Web search and grounding
Web search is a Pro feature on both platforms. Without it, models answer from training knowledge alone. With it:
- ChatGPT uses OpenAI's web search tool.
- Gemini uses Google Search grounding. Grounded Gemini answers return citation links through a Google redirect service; Obsurfable resolves these to their real destination URLs before storing them, so citation and competitor analysis works on actual domains.
Plan limits
On Free, each prompt can be run once per platform, and your 50-prompt allowance counts distinct prompts rather than runs. Asking one prompt on both ChatGPT and Gemini spends one prompt of the allowance, not two.
On Pro and Agency, re-runs are unlimited on every live platform and web search is available. See Plans and features.
FAQ
Do I have to pick one platform?
No. You can ask the same prompt on both, and scheduled runs cover every platform you have enabled. Comparing them is the point.
Will my historical data still work?
Yes. Results captured before Gemini support are attributed to ChatGPT, and your existing visibility score history is preserved as the overall series.
Why does my overall score differ from both platform scores?
The overall score blends results from every platform in the run. If Gemini mentions you and ChatGPT does not, the overall score lands between the two.
When are Claude and Perplexity coming?
Both are on the roadmap and already modelled in the app, which is why they show as Coming soon rather than being hidden. They will appear in the platform picker as soon as their integrations ship.