Appearance
Analytics & Monitoring
Analytics show how prompts perform across providers, including cost, latency, usage trends, and system health. Use these tools to optimize spending, catch problems early, and understand how your team uses the platform.
Usage Dashboard
The usage dashboard gives you a real-time view of platform activity. You can track:
- Request volume — total API calls over a selected time range
- Token consumption — input and output tokens broken down by model
- Cost — accumulated spend with per-request cost averages
- Unique users — how many team members are actively sending prompts
Switch between hourly, daily, weekly, and monthly views to spot trends at different scales. Filter by subscription or provider to narrow the data.
Usage Alerts
VeriPrompt automatically warns you when usage patterns need attention:
- Approaching rate limits — notifies you before requests start getting throttled, so you can adjust traffic or upgrade your plan.
- Subscription expiration — advance notice before your subscription renews or expires, giving you time to act.
- High usage relative to quota — flags when consumption is trending toward your plan limits, helping you avoid unexpected overages.
Alerts appear in the dashboard and can be configured per subscription.
Usage Export
Download your usage data for external analysis, audits, or reporting:
- Export as CSV or JSON
- Filter by date range and subscription
- Includes request counts, token usage, cost breakdowns, and provider details
Use exports to feed data into spreadsheets, BI tools, or internal billing systems.
Provider Performance
Compare how each AI provider is performing with per-provider metrics:
| Metric | What it tells you |
|---|---|
| Average latency | How fast the provider responds to requests |
| Error rate | Percentage of requests that fail |
| Success rate | Percentage of requests completed successfully |
| Cost per request | Average spend per API call to that provider |
| Daily performance trends | How these metrics change over time |
Use this data to decide which providers to prioritize in your routing rules or to identify providers that are underperforming.
System Health Monitoring (Admin)
Administrators have access to a real-time health dashboard with traffic-light indicators (green / yellow / red) for each subsystem:
- Application health — memory usage, uptime, and process status
- Database health — connection pool utilization and query latency
- Provider health — circuit breaker status for each configured provider (see below)
- Gateway health — overall error rate, p95/p99 latency, and request throughput
When an indicator turns yellow or red, the dashboard shows actionable recommendations — for example, suggesting you check a provider's status page or review recent configuration changes.
Circuit Breaker
VeriPrompt includes automatic provider failover so your prompts keep working even when a provider has problems:
- A provider with a greater than 50% error rate is temporarily taken out of rotation.
- After a 5-minute cooldown, the system sends test requests to check whether the provider has recovered.
- If the provider is healthy again, it is automatically re-enabled — no manual intervention needed.
This self-healing behavior means transient provider outages do not require you to change settings or respond to alerts. The system handles failover and recovery on its own.
When it helps
- Optimizing cost — identify which prompts or providers are driving the most spend, then adjust routing or prompt design accordingly.
- Debugging slowdowns — use latency trends and provider performance data to pinpoint whether delays come from a specific provider or from increased traffic.
- Capacity planning — monitor usage growth over weeks and months to decide when to upgrade your plan or add providers.
- Incident response — the health dashboard and circuit breaker status tell you immediately whether an issue is on the VeriPrompt side or a provider side.
- Compliance and reporting — export usage data for internal audits, client billing, or regulatory reporting.