> For the complete documentation index, see [llms.txt](https://docs.quper.co/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.quper.co/vidura-ai/vidura-ai.md). # Vidura AI Overview Vidura is Quper's built-in AI engine. It monitors hundreds of signals across your cloud infrastructure in real time, performs root cause analysis on cost events, and generates ranked optimisation recommendations — so your team always knows what is happening and what to do about it. Vidura is not a chatbot layered on top of your data. It is the intelligence layer that powers everything Quper does: every anomaly alert, every recommendation, and every root cause explanation comes from Vidura's analysis. ## What Vidura does ### Real-time monitoring Vidura continuously ingests usage and billing signals from all connected cloud providers and data platforms. It processes these signals against learned baselines to detect deviations the moment they occur — not hours or days later. ### Anomaly detection Vidura's ML models build a baseline of expected behaviour for every service, account, region, and resource in your environment. When usage deviates from this baseline, Vidura generates an alert with: * The specific resource and service causing the anomaly * The magnitude of the deviation (e.g. 3.4× normal usage rate) * The per-hour cost impact * A projected end-of-day cost if left unresolved ### Root cause analysis Every anomaly Vidura detects comes with an explanation in plain language. Vidura identifies what changed — a new resource launched without auto-stop, a query pattern that changed, a warehouse that stopped auto-suspending — and surfaces it alongside the alert so your team can act without manual investigation. ### Optimisation recommendations Vidura continuously scans your infrastructure for rightsizing opportunities, idle resources, missing configurations, and commitment gaps. Recommendations are ranked by monthly savings potential and include: * The specific action to take * The estimated monthly savings * A confidence level based on how long the pattern has been observed * One-click application for supported platforms ### Forecast and run rate Vidura projects your end-of-month spend based on current usage trajectory. If a cost anomaly is active, Vidura shows a separate "if unresolved" forecast so you can quantify the impact of inaction. ## How Vidura surfaces its analysis Vidura's outputs appear throughout the Quper interface: | Where | What Vidura shows | | ------------------------ | ------------------------------------------------------------ | | **Spend Overview** | Run rate forecast, anomaly markers on the spend chart | | **Anomalies feed** | Detected anomalies with root cause explanations | | **Action Intelligence** | Ranked optimisation recommendations with savings estimates | | **Governance dashboard** | Policy compliance analysis and gap identification | | **Integration setup** | Initial analysis and first recommendations on data ingestion | ## Vidura's learning period Vidura requires a minimum of **14 days of usage data** to establish reliable baselines for anomaly detection. During this period: * Spend Overview and recommendations are available immediately * Anomaly detection operates at reduced sensitivity with a wider deviation threshold * After 14 days, sensitivity increases and false positive rates decrease significantly Vidura continues to improve its baselines over time as it observes more of your infrastructure's behaviour patterns. ## Data privacy and security Vidura's analysis runs entirely within Quper's secure infrastructure. Your billing and usage data is never shared with third-party AI providers or used to train models outside of your workspace. Vidura's models are trained on your organisation's own cost and usage patterns only. Quper is SOC 2 Type II certified. All data is encrypted in transit and at rest. ## FAQs **Does Vidura have access to my actual workloads or data?** No. Vidura only analyses billing and usage metadata — the same data you see in your cloud provider's billing console. It has no access to the contents of your databases, code, or application data. **Can Vidura make changes to my cloud resources automatically?** Vidura can apply approved optimisations with one-click confirmation in Quper. It does not make autonomous changes to your infrastructure without explicit approval. All actions are logged in the Decision Trail. **How does Vidura compare to my cloud provider's native cost tools?** Native tools (AWS Cost Explorer, GCP Cost Management, Azure Cost Analysis) show you what happened. Vidura explains why it happened and what to do about it — across all your providers in a single place. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://docs.quper.co/vidura-ai/vidura-ai.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.