Appearance
Cinematic Mode: Interactive Visuals
VeriPrompt's routing, cost, security, and dependency data is real-time and detailed — but a table of numbers doesn't always make it obvious what's actually happening. Cinematic mode turns the platform's key screens into animated, interactive visuals that make routing decisions, spend flows, and risk exposure visible at a glance, instead of something you have to read row by row.
Every visual is built on real data from the same APIs that power the tables and reports next to it — nothing is simulated or decorative. If a number appears on screen, it came from your company's actual activity.
Turning it on or off
Cinematic mode is a per-user preference, so switching it does not affect your teammates.
- Go to Settings → Display.
- Toggle Cinematic mode on or off.
When it's off, every visual on the platform falls back to a plain, static view of the same data — nothing is hidden, it just stops animating.
Per-visual mute: even with Cinematic mode on, you can turn off a single visual you find distracting without affecting the rest. Look for the eye icon in the corner of any visual panel — click it to hide that one visual (it becomes the plain fallback), and click it again to bring it back.
Reduced motion: if your operating system is set to reduce motion (macOS: Reduce motion; Windows: Show animations, off), VeriPrompt automatically shows the static fallback everywhere, regardless of your Cinematic mode setting.
What you'll see
Below is a tour of the interactive visuals shipped so far, grouped by where you'll find them.
Routing & sending prompts
Policy Globe — /routing/policies, /admin/routing-policies A world globe showing your company's location and every AI provider's datacenter as a point on the map. Editing a routing policy re-simulates it live: a line launches from your company toward each candidate provider — blue if the route is allowed, red if it's blocked by the policy (the line stops exactly at the geofence boundary). Hover any provider for the rule that decided its fate. Use case: you just added a new geofencing rule and want to confirm it actually blocks the providers you intended, before saving.
Launch Sequence — Guru "Run It", Studio test runs Sending a prompt is shown as a round trip on the same globe: your prompt "launches" toward the chosen provider (passing visibly through the Shield sanitization ring if PII is detected), then the response "returns" and lands back home. A failed provider shows a deflected line and a second launch toward the fallback, with the reason. Use case: a request feels slow, and you want to see whether it failed over to a fallback provider along the way.
The Eleven Gates — Guru "Run It" result panel Every request's real journey through the gateway's eleven processing stages (security check, routing, sanitization, injection scan, provider call, response processing, and more), drawn as a corridor of gates that open one after another at their real recorded duration. A gate that failed slams shut in red instead of opening. Use case: a request took longer than expected, and you want to see exactly which stage of the pipeline was the bottleneck.
Token Swap — Shield's sanitization diff view (/shield/sanitize) Each piece of personal data detected in your document visibly scrambles character-by-character into its protective token (e.g. [PERSON_1]) and locks into place. A toggle lets you replay it in reverse to watch the original value restore. Use case: you want to show a colleague exactly what leaves your document (and what comes back) before trusting the sanitization pipeline.
First Flight — the BYOK onboarding wizard (/onboarding/byok) The first time you successfully test a newly-added API key, VeriPrompt plays the full Launch Sequence animation full-screen as a small "it works" moment, with a live caption for each real pipeline stage. Use case: you just connected your own OpenAI/Anthropic key and want confirmation it's actually wired up correctly.
Shield (data protection)
PII Sonar — the pre-send preview and /shield/sanitize A radar-style sweep reveals the categories of personal data detected in your content — one blip per category (persons, emails, phone numbers, IBANs, etc.), sized by how many were found, settling into a closed lock once sanitization is ready. Use case: before sending a document to an external model, you want a quick visual gut-check of what kinds of sensitive data it actually contains.
Sherlock (security & cost intelligence)
Attack Map — /sherlock/security A globe showing every inbound request VeriPrompt's IP security layer evaluated in the last 24 hours, as an arc from its real geolocated origin toward your company. Blocked requests shatter exactly at your geofence boundary; VPN/proxy/Tor traffic gets a distinct marker. Use case: you want to see, at a glance, whether the requests hitting your gateway are coming from where you'd expect.
Prompt-Injection Ambush — /sherlock/security A companion globe to Attack Map, focused specifically on detected prompt-injection attempts: an arc from the attacking source shatters at your defenses, with a burst of labeled attack-technique paths fanning out from the impact point. Use case: you want to understand not just that an injection attempt happened, but which techniques it tried.
Money River Sankey — /sherlock A flow diagram of where your AI spend actually goes: total spend → business unit → provider → model, drawn as flowing bands whose width is the real dollar amount. Use case: your finance lead asks "where does our AI budget go" — this answers it in one glance instead of a spreadsheet pivot.
Re-route Sankey — /sherlock/simulator Run a what-if scenario (e.g. "what if Marketing switched to Anthropic?") and watch the spend bands physically re-route from your current setup to the simulated one, with a companion chart projecting the cost difference over multiple future horizons (1 month, 1 year, etc.). Use case: you're evaluating a provider switch and want to see the projected savings before committing.
Blast Radius — /sherlock/dependencies An interactive map of which workflows depend on which providers. Click a provider to see its "blast radius" — the workflows that would be affected if it went down, split into full outage (their only provider) vs. partial exposure (they have another option) — with a live counter of the real dollar spend at risk. Use case: before renewing (or dropping) a provider contract, you want to know exactly which workflows would break.
Organisation Sunburst — /sherlock/org Your business-unit hierarchy drawn as concentric rings, where ring depth is the org structure and each ring's width is real spend (including everything underneath it). Click a ring to zoom into that unit. Use case: you want to drill from company-wide spend down to a specific team's number without leaving the chart.
Budget Breath — /sherlock/budgets Every active budget shown as a "vessel" that fills to its real utilization. It breathes gently while healthy, breathes faster once it passes its warning threshold, and freezes solid — with a lock icon — the moment a hard-capped budget is actually blocking new spend. Use case: at a glance, you want to know which budgets need attention before checking each one individually.
Waste Ledger — /sherlock/findings Detected waste and risk findings (idle subscriptions, redundant credentials, single-provider dependencies, and more) shown as droplet-shaped "leaks," sized by estimated waste and colored by severity. Acknowledging or resolving a finding seals its leak, and a recovered-savings counter animates upward. Use case: you're triaging a list of cost-optimization findings and want to feel the visible impact of clearing them.
Accessibility
Every visual respects your system's reduced-motion setting automatically, and every visual can be individually muted without affecting Cinematic mode for the rest of the platform. Turning Cinematic mode off entirely never hides functionality — it only removes the animation, leaving the same real data in a plain, static view.