FirstData

FirstData aims to build a global, authoritative, structured 'external facts context layer' for AI agents by aggregating trustworthy sources across research, government disclosures, laws and regulations, corporate filings, and industry standards, turning scattered raw material into traceable, citable core facts with a full evidence chain and version history. Every listed source is classified into one of several authority tiers, such as government, international organization, research institution, market, and commercial, and every URL is stated to be verified so links do not silently break.

Each source's metadata is bilingual, in Chinese and English, and includes its access URL, API endpoint if any, download method, authority tier, and update frequency, structured so it can be queried programmatically for automated evidence tracing rather than only browsed as a list of links. The project ships a standard MCP server so AI applications can query the catalog directly, and an Agent Skill definition so an agent can register, activate, and configure the MCP connection itself without manual setup. The project states a priority on covering Chinese government data sources, alongside international organizations such as the World Bank, IMF, OECD, WHO, and FAO.