> ## Documentation Index > Fetch the complete documentation index at: https://docs.heralded.ai/llms.txt > Use this file to discover all available pages before exploring further. # Snapshot > The one-off report on one brand: what it asks, what it shows, and how it carries into the Command Center. A Snapshot is a one-off report on one brand. It asks the engines your buyers' questions once, reads the answers the way the Command Center does, and gives you a private report. Your first Snapshot is free, within a daily allowance of free Snapshots: when today's are taken, yours is saved and you can start it after midnight UTC. Later ones use a Snapshot credit, priced on [heralded.ai/snapshot](https://heralded.ai/snapshot). ## Starting a Snapshot Start your first Snapshot at [heralded.ai/snapshot](https://heralded.ai/snapshot). Signed in, use **New Snapshot** to enter a domain and **My Snapshots** to open your reports at [app.heralded.ai/snapshots](https://app.heralded.ai/snapshots). Your Snapshots open inside the Command Center. Subscribers reach **Snapshots** from the account menu. Without a plan, Snapshots stay open and the other sections are locked. Choose a locked section to start your free seven-day trial without a card. Shared reports and their briefs open on their own. Review the brief before choosing **Run Snapshot →**. It says how many Snapshots the run uses. When your balance is zero, **Buy more** opens the Snapshot bundles. Buying from a brief returns you to that brief. Your balance updates when payment clears. Canceling checkout also returns you to the brief. ## What it asks * 20 prompts, drafted from your site. You review them, with your brand profile and competitors, before it starts. * The six standard [engines](/concepts/engines). Claude is part of the Command Center only. * Each prompt twice on each engine, so 240 answers. * Up to five competitors, in one market. ## Which of your pages it reads A Snapshot reads up to 150 pages of your site. Heralded lists your site's pages, sorts them by type, and reads pricing, solution and product pages first, then the rest of your content. It keeps to your site's language. Your homepage counts toward this limit. Pages found through topic searches or engine citations use the remaining slots. Community threads do not take priority slots. Another Snapshot of the same site started in the same week reads the pages the first one read. A re-scan reads your site again. Where buyers ask about something no page of yours covers, the Snapshot names your top five content ideas, in the report and in its Audit. The Command Center carries the rest. ## What the report shows The [Heralded Score](/methodology/heralded-score) with [Mentioned and Recommended](/methodology/mentioned-and-recommended), for all engines together, for each engine above your prompts and for each prompt, [how AI describes you](/methodology/how-ai-describes-you), how you compare with your competitors, the [sources](/concepts/sources) the engines cite, [what stops AI engines reading your site](/concepts/site-readability), and the moves worth making first. Every claim links back to the answers behind it, and the report also comes as a PDF. The figures mean what [How Heralded measures](/methodology/how-we-measure) says. [Change](/methodology/change) and [verifying actions](/methodology/verifying-actions) need weekly reports, so a Snapshot has neither. Your report opens and its email goes out as soon as scoring finishes. Analysis, your content and share cards follow. If analysis is still being written when you open your report, its summary, positioning insights, rival framing and priority actions say so. Your content also shows when it is still being prepared. Each section fills in automatically as it finishes. Your measured results stay available while you wait. Each share card shows a loading state until its preview is ready. You can download it once it appears, without reloading the report. You can generate the PDF once the report is final and its analysis and content sections have finished. If either section fails, the PDF includes its unavailable state. It waits for any final content update so your PDF captures the finished report. ## When an engine is slow Heralded starts scoring as soon as all answers are in, and waits up to 12 minutes for a slow engine. It reads cited pages while the engines answer, so those pages are ready for scoring. After 12 minutes, Heralded can start scoring with fewer answers if enough engines have answered. When coverage is too low, it checks again every five minutes, up to a 30-minute limit. If at most 5% of the answers are still missing, and no engine is missing more than 20% of its own, those answers are [not measured](/methodology/how-we-measure#which-answers-count) and the report opens whole. With more missing, the report opens with partial data and the email tells you so. Missing answers keep arriving, and the report fills in as they do. If the remaining answers arrive before delivery, Heralded scores them and brings the analysis up to date before opening one whole report and sending one email. A partial report says so at the top. The positioning map, AI understanding, your content and the priority actions are marked **Updates when the report is final**. They are written again once it is. While Heralded is still processing answers or finalizing the report, a report you have open fills in on its own, without a reload. The report is final once every answer is in, or 105 minutes after the partial report went out, whichever comes first. If finalization is delayed, the report stays partial while Heralded retries it. If Heralded cannot finalize it, your partial report remains available with the measurements it already contains. Answers that are still missing are [not measured](/methodology/how-we-measure#which-answers-count): they're left out of the figures and don't count against you. Heralded emails you when the report is final. While the report is partial, its PDF, share cards and next Snapshot wait. Its link preview has no score. Your account's report list shows no score until the report is final. If a section cannot be displayed, the rest of your report stays available. Reload the page to try loading the section again. The section also retries when the report updates. ## Running it again Running a Snapshot again is a fresh run. It asks the engines again, keeps its answers apart from the earlier report's, and uses the same prompts so the two compare like with like. Reports on score version v7.0 open when they have a usable frozen snapshot. Older reports and reports without a usable frozen snapshot do not open, and your account's report list leaves them out. Start a new Snapshot to get a report on the current scale. Once a new Snapshot is delivered, you run it again from the newest report on the link. Reports from Heralded's pre-launch period have been retired; their links say so and offer a fresh Snapshot. ## In the API Read a Snapshot's Heralded Score, Mentioned, Recommended and readiness through `GET /v1/scans/{scan_id}` or MCP `get_scan`, and its report as a link holder sees it through `GET /v1/shared/scans/{share_id}`. [Scans in the API](/figures#scans) explains the fields. ## Moving to the Command Center When you subscribe, a Snapshot you already bought becomes the brand's starting point. Its prompts start ticked during [setup](/get-started), and it is scored again over the prompts you track, as the first point on your trend. If you subscribe while your Snapshot is partial, the dashboard shows it as a read in progress. Setup uses it as your starting point once the report is final. If it cannot be finalized, setup starts a fresh first read instead. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.