> ## Documentation Index
> Fetch the complete documentation index at: https://docs.wardin.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# What is Wardin?

> A vendor-neutral AI Governance & Observability platform that sits in the request path.

Wardin is a **policy-enforcing LLM gateway** combined with **finance-grade cost allocation** and **productivity attribution tied to business outcomes** — three categories that are normally three separate vendor products, built as one system.

## The core problem

Most teams deploying AI at scale stitch together a gateway + observability tool + guardrail service + FinOps tool + productivity tool. Each sits outside the request path, producing dashboards that lag reality, and none of them enforces anything.

Wardin sits **in the request path** — between your code and every LLM provider. That means:

* Budgets are enforced before a token is spent, not discovered in a report
* PII is redacted before it leaves your network
* Policies (model allowlist, injection detection) block non-compliant calls at the source
* Every request is attributed to a team, user, and session for ROI tracking

## How it works

```mermaid theme={null}
flowchart LR
    A([Your code]) -->|wardin_sk_| GW[Wardin Gateway]
    GW --> B[Auth]
    B --> C[Budget check\nRedis atomic]
    C --> D[Policy enforcement\nmodel · guardrail · PII]
    D --> E[Cache lookup\nL1 exact · L2 semantic]
    E -->|cache miss| F{Provider}
    F --> G([Anthropic])
    F --> H([OpenAI])
    F --> I([Bedrock / Gemini])
    E -->|cache hit| R([Response])
    G & H & I --> J[Parse token costs]
    J --> K[Commit spend]
    K --> L[Async emit\nClickHouse]
    K --> R
```

You point your existing SDK at `https://gw.wardin.ai` instead of `https://api.anthropic.com`. No client code changes.

## Key concepts

<CardGroup cols={2}>
  <Card title="Virtual Keys" icon="key" href="/concepts/virtual-keys">
    Scoped API keys that proxy to your real provider credentials — with per-key budgets, rate limits, and policies.
  </Card>

  <Card title="Budgets" icon="gauge" href="/concepts/budgets">
    Hard spending limits enforced via Redis atomic counters before a request is forwarded.
  </Card>

  <Card title="Policies" icon="shield" href="/concepts/policies">
    Model allowlists, prompt injection detection, PII redaction — enforced in-path.
  </Card>

  <Card title="Caching" icon="bolt" href="/concepts/caching">
    L1 exact-match and L2 semantic caching reduce spend without changing your prompts.
  </Card>

  <Card title="Sessions" icon="layer-group" href="/concepts/sessions">
    Requests grouped into sessions so cost-per-task is a real metric — built for agentic clients.
  </Card>

  <Card title="Signed Receipts" icon="file-signature" href="/concepts/receipts">
    Every request emits an ED25519-signed, hash-chained receipt — tamper-evident audit evidence.
  </Card>
</CardGroup>