docs: add dashboard screenshot and update docs to mention /dashboard

- README: new "Monitoring Dashboard" section with screenshot and feature list,
  dashboard added to API endpoint table, roadmap item marked done
- QUICKSTART: dashboard step added after statistics curl examples
- PROJECT_OVERVIEW: /dashboard row added to API endpoints table

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: 2 people

PROJECT_OVERVIEW.md

@@ -217,6 +217,7 @@ FastAPI application that wraps the pipeline.
 | `/api/v1/config` | GET | — | Sanitised safety config |
 | `/health` | GET | — | Server health + uptime |
 | `/health/backend` | GET | — | Remote LLM reachability probe |
+| `/dashboard` | GET | — | Built-in monitoring dashboard UI |
 
 **Environment variables:**
 

QUICKSTART.md

@@ -244,6 +244,15 @@ curl -X POST http://localhost:8000/api/v1/statistics/reset \
   -H "X-API-Key: <ADMIN_API_KEY>"
 ```
 
+#### 6. Monitoring dashboard (browser)
+
+Open **http://localhost:8000/dashboard** in your browser for a live visual
+overview of all the above — no extra tools needed. The dashboard polls the API
+every 10 seconds and shows request counts, accept rate, rejection reasons,
+latency, and backend health.
+
+![Dashboard preview](docs/dashboard-preview.png)
+
 ---
 
 ## 📊 Common Use Cases

README.md

@@ -395,11 +395,30 @@ python api_server.py
 | `GET /api/v1/config` | Sanitised safety config |
 | `GET /health` | Server health + uptime |
 | `GET /health/backend` | Remote LLM reachability probe |
+| `GET /dashboard` | Monitoring dashboard UI |
 
 Interactive docs: `http://localhost:8000/docs`
 
 ---
 
+## Monitoring Dashboard
+
+A built-in real-time dashboard is served at **`http://localhost:8000/dashboard`** — no extra tooling required.
+
+![LLM Safety Dashboard](docs/dashboard-preview.png)
+
+The dashboard auto-refreshes every 10 seconds and shows:
+
+- **Live status chips** — server health and backend reachability
+- **Request counts** — total / accepted / rejected with percentages
+- **Average latency** per request
+- **Accept rate** progress bar
+- **Accept / reject donut chart** (pure SVG, no external dependencies)
+- **Rejection reason breakdown** — colour-coded bars for each rejection type (`token_filter`, `pattern_match`, `semantic_unsafe`, etc.)
+- **System info** — version, uptime, backend type, model, backend URL
+
+---
+
 ## Testing
 
 ```bash
@@ -450,7 +469,7 @@ docker run -p 8000:8000 \
 ## Roadmap
 
 - [ ] Multi-language support (non-English prompts)
-- [ ] Real-time monitoring dashboard
+- [x] Real-time monitoring dashboard (`/dashboard`)
 - [ ] Streaming response support
 - [ ] Custom classifier model plug-in API
 - [ ] Distributed deployment with shared rate-limit store (Redis)