Put SocialAI Studio's AI tools inside your own apps.

1 · Initialize

curl -s https://socialai.website/api/mcp \
-H "Authorization: Bearer sk_socialai_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "initialize",
    "params": {
      "protocolVersion": "2025-03-26",
      "capabilities": {},
      "clientInfo": { "name": "my-app", "version": "1.0.0" }
    }
  }'

2 · List tools

curl -s https://socialai.website/api/mcp \
-H "Authorization: Bearer sk_socialai_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{
    "jsonrpc": "2.0",
    "id": 2,
    "method": "tools/list",
    "params": {}
  }'

3 · Call a tool (credit balance)

curl -s https://socialai.website/api/mcp \
-H "Authorization: Bearer sk_socialai_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{
    "jsonrpc": "2.0",
    "id": 3,
    "method": "tools/call",
    "params": { "name": "get_credit_balance", "arguments": {} }
  }'

4 · Generate a video, then poll for it

curl -s https://socialai.website/api/mcp \
-H "Authorization: Bearer sk_socialai_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{
    "jsonrpc": "2.0",
    "id": 4,
    "method": "tools/call",
    "params": {
      "name": "generate_video",
      "arguments": {
        "prompt": "A golden retriever running through a field at sunset, cinematic",
        "duration": 10,
        "quality": "fast",
        "ratio": "16:9"
      }
    }
  }'
# Returns {"content":[{"type":"text","text":"job_id: ..."}]}
# Then poll until it is done:
curl -s https://socialai.website/api/mcp \
-H "Authorization: Bearer sk_socialai_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{
    "jsonrpc": "2.0",
    "id": 5,
    "method": "tools/call",
    "params": {
      "name": "check_job_status",
      "arguments": { "job_id": "PASTE_JOB_ID" }
    }
  }'

5 · Generate a video from a photo (image-to-video, face locked)

curl -s https://socialai.website/api/mcp \
-H "Authorization: Bearer sk_socialai_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{
    "jsonrpc": "2.0",
    "id": 6,
    "method": "tools/call",
    "params": {
      "name": "generate_video",
      "arguments": {
        "prompt": "The founder speaks to camera about her skincare brand, natural lip-sync, steady camera",
        "duration": 10,
        "quality": "fast",
        "ratio": "9:16",
        "image_url": "https://example.com/founder-photo.jpg",
        "privacy_consent": true
      }
    }
  }'
# The video starts from the exact photo: the photo is sent as the
# first frame AND as a character reference, and the best face-likeness
# model is picked automatically, so the person in the video is the
# person in the photo — not a generic lookalike. privacy_consent=true
# confirms you are the person in the photo or have their written
# permission; it is REQUIRED for photo jobs and is stored with the
# job. Without it, a photo job is refused. image_data_url
# (base64 PNG/JPEG/WebP) works the same way; provide only one of the
# two. generate_long_video takes the same two fields, and EVERY
# segment starts from the same photo. Image-to-video supports up to
# 30s per clip / segment: use segment_duration 10 or 30 for long
# videos with a photo.

Claude Code — one command (copy-paste)

claude mcp add --transport http socialai https://socialai.website/api/mcp --header "Authorization: Bearer <your-api-key>"

Verify it connected

claude mcp list
# You should see: socialai — connected

Claude Code (.mcp.json) / Claude Desktop / Cursor (JSON config)

{
  "mcpServers": {
    "socialai": {
      "type": "http",
      "url": "https://socialai.website/api/mcp",
      "headers": { "Authorization": "Bearer <your-api-key>" }
    }
  }
}