Presenton Agent Skill

# Docker (Linux/macOS) — recommended
docker run -it --name presenton \
  -p 5000:80 \
  -v "./app_data:/app_data" \
  ghcr.io/presenton/presenton:latest

# Docker (Windows PowerShell)
docker run -it --name presenton `
  -p 5000:80 `
  -v "${PWD}\app_data:/app_data" `
  ghcr.io/presenton/presenton:latest

# With OpenAI + DALL-E 3 (no UI key entry needed)
docker run -it --name presenton \
  -p 5000:80 \
  -e LLM="openai" \
  -e OPENAI_API_KEY="<your-key>" \
  -e IMAGE_PROVIDER="dall-e-3" \
  -e CAN_CHANGE_KEYS="false" \
  -v "./app_data:/app_data" \
  ghcr.io/presenton/presenton:latest

# Start Presenton first (see install above), then call the API
curl -fsS --max-time 60 \
  -X POST "http://localhost:5000/api/v1/ppt/generate" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Introduction to Quantum Computing", "n_slides": 8}'

# The response includes a presentation ID; download PPTX with:
PPTX_PATH=$(curl -s "http://localhost:5000/api/v1/ppt/generate" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Introduction to Quantum Computing", "n_slides": 8}' \
  | jq -r '.pptx_url')
curl -s "http://localhost:5000${PPTX_PATH}" -o presentation.pptx

async function generatePresentation(prompt, nSlides = 8, baseUrl = 'http://localhost:5000') {
  const res = await fetch(`${baseUrl}/api/v1/ppt/generate`, {
    method: 'POST',
    headers: { 'Content-Type': 'application/json' },
    body: JSON.stringify({ prompt, n_slides: nSlides }),
  });
  if (!res.ok) throw new Error(`HTTP ${res.status}: ${await res.text()}`);
  return await res.json(); // contains pptx_url and pdf_url
}

// Usage
// generatePresentation('Introduction to Quantum Computing', 10).then(console.log);

# Add Presenton MCP server to your agent config (e.g. Claude Desktop, Cursor)
# mcp.json entry:
cat <<'EOF'
{
  "mcpServers": {
    "presenton": {
      "url": "http://localhost:5000/mcp"
    }
  }
}
EOF

// The MCP server exposes a generate_presentation tool.
// Call it via your MCP client library:
const result = await mcpClient.callTool('presenton', 'generate_presentation', {
  prompt: 'Climate Change: Causes and Solutions',
  n_slides: 10,
});
console.log(result); // { pptx_url, pdf_url }

# Upload a template PPTX to extract theme/design
curl -fsS --max-time 30 \
  -X POST "http://localhost:5000/api/v1/ppt/upload-template" \
  -F "file=@my_template.pptx"

# Generate a new presentation using that template
curl -fsS --max-time 60 \
  -X POST "http://localhost:5000/api/v1/ppt/generate" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Q3 Sales Report", "n_slides": 6, "template": "my_template"}'

const { readFileSync } = require('fs');

async function uploadTemplate(filePath, baseUrl = 'http://localhost:5000') {
  const form = new FormData();
  form.append('file', new Blob([readFileSync(filePath)]), 'template.pptx');
  const res = await fetch(`${baseUrl}/api/v1/ppt/upload-template`, {
    method: 'POST',
    body: form,
  });
  if (!res.ok) throw new Error(`Upload failed: HTTP ${res.status}`);
  return await res.json();
}

// Usage
// uploadTemplate('./branding.pptx').then(console.log);

You have Presenton capability. When a user asks to create a presentation:

1. Confirm the topic and desired number of slides (default: 8)
2. Call POST http://localhost:5000/api/v1/ppt/generate with {"prompt": "<topic>", "n_slides": <n>}
3. Wait for the response (up to 60 seconds) and extract pptx_url and pdf_url
4. Offer the user download links for both PPTX and PDF
5. If a custom template is requested, upload it first via /api/v1/ppt/upload-template and include the template name in the generate request

Always check that Presenton is running at http://localhost:5000 before calling the API.
Report any HTTP errors with the status code and response body so the user can diagnose the issue.