KrillData: live web and social data for AI agents over MCP

# KrillData > Live web and social data for AI agents: 519 endpoints over 127 sources, one API key, one JSON schema. ## MCP Remote MCP server: https://data.openkrill.app/mcp, with the API key in the `x-api-key` header. Tools: `search_endpoints` finds an endpoint; `call_endpoint` calls it by `path` with `params`. Claude Code (Run in a terminal): ``` claude mcp add --transport http krilldata \ https://data.openkrill.app/mcp \ --header "x-api-key: YOUR_API_KEY" ``` Cursor (Add to ~/.cursor/mcp.json): ``` { "mcpServers": { "krilldata": { "url": "https://data.openkrill.app/mcp", "headers": { "x-api-key": "YOUR_API_KEY" } } } } ``` Claude Desktop (Add to claude_desktop_config.json): ``` { "mcpServers": { "krilldata": { "command": "npx", "args": [ "-y", "mcp-remote", "https://data.openkrill.app/mcp", "--header", "x-api-key:${API_KEY}" ], "env": { "API_KEY": "YOUR_API_KEY" } } } } ``` ## Base URL https://data.openkrill.app/v1 ## Authentication Send the key in the `x-api-key` header, or as `Authorization: Bearer <key>`. Never put the key in a query string. ## Response envelope Success: `{ success, platform, endpoint, data, credits_used, credits_remaining, request_id, cached }`. Failure: `{ success: false, error: { type, message, status, doc_url, retryable }, request_id }`. `data` always contains `resources[]`, and each resource has `author`, `engagement`, `metadata`, and `extras`. Parse those four keys once and every platform works. ## Credits - Each endpoint costs 1 or 5 credits per live call; its docs page states which. - Cache hits cost 0 credits and return `cached: true` with `X-Cache: HIT`. ## Caching Freshness windows follow the data type: profiles 15m, posts 10m, comments 5m, search 2m, analytics 30m. Send `Cache-Control: no-cache` to force a live fetch at the normal price. If an upstream refuses while you still hold a