Unifapi | Give Your Agent Live Public Data
UnifAPI gives Codex, Claude Code, Cursor, and other MCP-capable agents live public data for task-specific workflows. Start from a Skill such as KOL pricing, connect the hosted MCP server when the agent needs data, and use the HTTP API reference when you want to turn a working workflow into product code.
What is UnifAPI? [#what-is-unifapi]
UnifAPI is a Skills-first public-data layer for AI agents. The API catalog is the data layer under those Skills: normalized public social records, consistent authentication, canonical response schemas, cursor pagination, structured errors, and OpenAPI discovery.
Who should use UnifAPI? [#who-should-use-unifapi]
Use UnifAPI when an agent or application needs public social data without building a separate integration for every provider. Common workflows include KOL pricing, creator research, social listening, competitor monitoring, Reddit community research, and China social trend scouting.
How is UnifAPI different from an API marketplace? [#how-is-unifapi-different-from-an-api-marketplace]
API marketplaces list many unrelated providers with different auth patterns, schemas, pagination rules, errors, and billing models. UnifAPI starts from agent tasks, then exposes the live public-data operations those tasks need through MCP and HTTP with one shape, one workspace, and one bill.
Get started [#get-started]
How it works [#how-it-works]
Why UnifAPI [#why-unifapi]
- Agent-native. MCP exposes
list_operations,get_operation, andcall_apiso an agent can discover and use the live catalog. - Public-data scoped. UnifAPI serves public records. OAuth authorizes the UnifAPI workspace, not a user's private upstream account.
- Discoverable.
/openapi.json,/llms.txt,/llms-full.txt, and markdown page alternates are first-class. - Portable. Skills can run in Codex or Claude Code; developers can later use the same operations over HTTP.
Server Config
{
"mcpServers": {
"unifapi": {
"url": "https://mcp.unifapi.com"
}
}
}