docs: add API Reference with examples, update Home with Mistral branding

Author: kiku-jw

wiki/API-Reference.md

@@ -0,0 +1,125 @@
+# Chart2CSV API Reference
+
+> **Base URL:** `https://chart2csv.kikuai.dev`
+> 
+> **Powered by Mistral Pixtral** — 90%+ accuracy on chart extraction
+
+---
+
+## Endpoints
+
+### `POST /extract`
+
+Extract data from a chart image.
+
+#### Parameters
+
+| Parameter | Type | Default | Description |
+|-----------|------|---------|-------------|
+| `file` | file | required | Chart image (PNG, JPG, WebP) |
+| `mode` | string | `llm` | Extraction mode: `llm`, `cv`, `auto` |
+| `chart_type` | string | auto | Force type: `scatter`, `line`, `bar` |
+
+#### Modes
+
+- **`llm`** (default) — Mistral Pixtral vision extraction, 90%+ accuracy
+- **`cv`** — Computer vision pipeline (legacy), faster but less accurate
+- **`auto`** — Try LLM first, fall back to CV if it fails
+
+---
+
+## Examples
+
+### curl
+
+```bash
+curl -X POST "https://chart2csv.kikuai.dev/extract" \
+  -F "file=@chart.png" \
+  -F "mode=llm"
+```
+
+### Python
+
+```python
+import requests
+
+url = "https://chart2csv.kikuai.dev/extract"
+
+with open("chart.png", "rb") as f:
+    response = requests.post(
+        url,
+        files={"file": f},
+        data={"mode": "llm"}
+    )
+
+result = response.json()
+print(f"Extracted {len(result['data'])} points")
+print(f"CSV:\n{result['csv']}")
+```
+
+### JavaScript (fetch)
+
+```javascript
+const form = new FormData();
+form.append('file', fileInput.files[0]);
+form.append('mode', 'llm');
+
+const response = await fetch('https://chart2csv.kikuai.dev/extract', {
+  method: 'POST',
+  body: form
+});
+
+const result = await response.json();
+console.log(`Extracted ${result.data.length} points`);
+console.log('CSV:', result.csv);
+```
+
+### JavaScript (axios)
+
+```javascript
+import axios from 'axios';
+
+const form = new FormData();
+form.append('file', file);
+form.append('mode', 'llm');
+
+const { data } = await axios.post(
+  'https://chart2csv.kikuai.dev/extract',
+  form
+);
+
+console.log(data.csv);
+```
+
+---
+
+## Response Format
+
+```json
+{
+  "success": true,
+  "chart_type": "scatter",
+  "confidence": 0.95,
+  "data": [
+    {"x": 0, "y": 10},
+    {"x": 1, "y": 20},
+    {"x": 2, "y": 30}
+  ],
+  "csv": "x,y\n0,10\n1,20\n2,30",
+  "warnings": [],
+  "processing_time_ms": 2500
+}
+```
+
+---
+
+## Rate Limits
+
+- **20 requests per minute** per IP
+- Contact us for higher limits
+
+---
+
+## Swagger UI
+
+Interactive API documentation: [chart2csv.kikuai.dev/docs](https://chart2csv.kikuai.dev/docs)

wiki/Home.md

@@ -1,37 +1,52 @@
-# Chart2CSV Wiki
+# Chart2CSV
 
-Welcome to the Chart2CSV documentation.
+> **Extract data from chart images using AI. 90%+ accuracy with Mistral Pixtral.**
 
-## What is Chart2CSV?
-
-Chart2CSV is a tool that extracts data from chart images and converts it to CSV format. You upload a picture of a chart (like a line graph or bar chart from a PDF or screenshot), and Chart2CSV gives you the actual numbers.
+---
 
 ## Quick Links
 
-- [Home](Home) — Overview
-- [Installation](Installation) — How to install
-- [Quick Start](Quick-Start) — Get started in 2 minutes
-- [CLI Reference](CLI-Reference) — All command line options
-- [How It Works](How-It-Works) — Technical explanation
-- [FAQ](FAQ) — Common questions
-- [Troubleshooting](Troubleshooting) — Fixing problems
+- 🚀 [Quick Start](Quick-Start)
+- 📖 [API Reference](API-Reference)
+- 💻 [CLI Reference](CLI-Reference)
+- ⚙️ [Installation](Installation)
+- 🔧 [Troubleshooting](Troubleshooting)
+- ❓ [FAQ](FAQ)
+
+---
+
+## Live Demo
 
-## Why Use Chart2CSV?
+Try it now: **[chart2csv.kikuai.dev](https://kiku-jw.github.io/Chart2CSV/)**
 
-**Problem:** You have a chart image (from a research paper, report, or website) and need the raw numbers.
+---
 
-**Old way:** Manually click each point in WebPlotDigitizer. Takes 5-30 minutes per chart.
+## Features
 
-**Chart2CSV way:** Drop the image, get CSV instantly. AI reads the chart for you.
+| Feature | Description |
+|---------|-------------|
+| 🧠 **Mistral Pixtral** | Powered by Mistral AI vision model |
+| ⚡ **Zero-Click** | AI reads charts automatically |
+| 📊 **Multi-Chart** | Line, scatter, bar charts |
+| 🔒 **Privacy** | Offline mode with Tesseract |
+| 🌐 **REST API** | Simple HTTP endpoint |
+
+---
 
 ## Supported Charts
 
-- ✅ Line charts
-- ✅ Scatter plots 
-- ✅ Bar charts
-- ✅ Linear and logarithmic scales
+✅ Line charts  
+✅ Scatter plots  
+✅ Bar charts  
+
+❌ Pie charts  
+❌ Heatmaps  
+❌ 3D charts  
+
+---
 
-## Getting Help
+## Technology
 
-- [GitHub Issues](https://github.com/kiku-jw/Chart2CSV/issues) — Report bugs
-- [FAQ](FAQ) — Common questions answered
+- **Mistral Pixtral Large** — Vision LLM for chart understanding
+- **FastAPI** — High-performance REST API
+- **OpenCV** — Image preprocessing (fallback mode)