Social media data for AI agents
An AI agent can read your social performance the same way a chat assistant can: by calling the MySoMeData API or MCP. MySoMeData is a read-only social media data API and MCP layer for AI. It connects social accounts to AI applications and software, providing structured historical performance data that can be analyzed, queried, and used to build products. The agent does not get a special feed. It gets the same normalized history.
How does an agent use the data?
You schedule the agent, or you let a person trigger it. The agent requests a period and a platform, receives JSON, and writes the analysis, the Slack note, or the next step in your product. Authentication is a Bearer key for HTTP, or the MCP session for a client that speaks MCP. The scopes stay read-only, so the agent cannot publish.
Which calls are enough?
- Start with the account performance summary for the strategy and the period comparison.
- Add historical trends when you need a daily series or a sample of items.
- Add outliers when the question is which items beat the usual volume.
- Add audience demographics only when the question is who, not what performed.
What should the agent be told?
Tell it the metric basis per platform, that item arrays are samples, and that one MySoMeData account is one company. Tell it to skip platforms the account did not connect. A weekly agent that re-fetches the summary is more reliable than one that caches a CSV from last quarter.