You don’t start collecting the day you connect.
You get the history those platforms make available, in one structure. Your AI does not have to learn three APIs. Engagement, reach, retention, and outliers are already calculated, and you read them through MCP or one API.
Engagement
Engagement you can trust
Ratios calculated across your posts, not an average of averages, so your AI works with the real rate.
Discovery
Reach outside your followers
See follower versus non-follower reach, plus Reels and TikTok watch-time, so you know when content leaves your existing audience.
Retention
Retention and traffic
See where viewers drop off, and whether traffic came from For You, Search, or Following, without parsing a raw TikTok response.
Momentum
Outliers and momentum
Spot posts well above your average, and see month-over-month change in followers, impressions, and engagement.
For AI
A payload your AI can read
One compact shape for context windows and automations. Same object whether you call the API or connect through MCP.
One API instead of three social APIs.
Your AI gets calculated metrics instead of raw platform responses. Generate a workspace-scoped key, then call REST or connect Claude, ChatGPT, n8n, or your own application through MCP. Rotate the key without reconnecting your accounts.
Base URL
https://api.mysomedata.com/v1
Auth
Bearer API key
Rate limit
120 requests / minute
Format
Normalized JSON, ready for AI
MCP
Streamable HTTP /api/mcp
Call the API directly, or give your AI access through MCP. Claude, ChatGPT, Gemini, n8n, Cursor, and your own app read the same data.
Your AI spends its context on analysis, not arithmetic.
One API and MCP. The same calculated data, whether you ask your AI or call it from a product.
Clean, historical, normalized social data.
API FAQ
Keys, payloads, and how normalized metrics map to connected accounts.
