# VidTranslate MCP Server

YouTube videos ko AI se translate karke naya video banao — Thesys Agent Builder ke liye MCP server.

## Pipeline
YouTube URL → Transcript → GPT-4 Translation → ElevenLabs Voice → D-ID Avatar Video → Publish

## Tools
| Tool | Kaam | API |
|------|------|-----|
| `get_transcript` | YouTube transcript extract | Free |
| `translate_script` | Text translation | OpenAI GPT-4 |
| `generate_voice` | AI voiceover | ElevenLabs |
| `create_video` | Avatar/Slideshow video | D-ID |
| `publish_video` | YouTube/FB/IG publish | Platform APIs |

---

## Railway Pe Deploy Karo (FREE)

### Step 1 — GitHub pe upload karo
```bash
git init
git add .
git commit -m "VidTranslate MCP Server"
git branch -M main
git remote add origin https://github.com/AAPKA-USERNAME/vidtranslate-mcp.git
git push -u origin main
```

### Step 2 — Railway pe deploy karo
1. railway.app pe jao — free account banao
2. "New Project" → "Deploy from GitHub repo" click karo
3. Apna `vidtranslate-mcp` repo select karo
4. Railway automatically deploy kar dega!
5. "Settings" → "Domains" → "Generate Domain" click karo
6. Aapka URL milega: `https://vidtranslate-mcp-XXXX.railway.app`

### Step 3 — Test karo
Browser mein kholo:
```
https://aapka-railway-url.railway.app/
https://aapka-railway-url.railway.app/health
https://aapka-railway-url.railway.app/mcp
```

---

## Thesys Mein Connect Karo

Thesys Agent Builder → Tools/MCP section mein:

```
MCP Server URL: https://aapka-railway-url.railway.app/mcp

Headers:
{
  "x-openai-key": "sk-YOUR-OPENAI-KEY",
  "x-elevenlabs-key": "YOUR-ELEVENLABS-KEY", 
  "x-did-key": "Basic YOUR-DID-KEY",
  "x-yt-token": "YOUR-YOUTUBE-TOKEN"
}
```

---

## API Keys Kahan Se Milenge

| API | Link | Free Tier |
|-----|------|-----------|
| OpenAI | platform.openai.com | $5 free credit |
| ElevenLabs | elevenlabs.io | 10k chars/month free |
| D-ID | studio.d-id.com | 20 free videos |
| YouTube | console.cloud.google.com | Free |

---

## Local Test Karo

```bash
pip install -r requirements.txt
uvicorn main:app --reload --port 8000
```

Browser: http://localhost:8000