Show HN: FactIQ – a realtime econ+finance database for AI agents
TL;DR
FactIQ is a new GitHub plugin from Defog AI that connects Claude Code and Codex to a finance and economics database for agent-driven research. The warehouse covers SEC filings, official statistics from the US, China, India, South Korea, IMF and World Bank, plus live market data and earnings-call intelligence. Agents can discover series, run read-only SQL, compute derived metrics and publish outputs as FactIQ charts, reports, terminal previews or local HTML visualizations.
Nauti's Take
FactIQ is not a flashy model launch. It is infrastructure, and that makes it interesting.
Research agents usually do not fail because their prose is weak; they fail because sources are messy, definitions shift and the first chart looks more certain than it should. The catch is trust.
Finance data needs freshness, caveats and clear provenance. A small repo making a Bloomberg-style comparison still has to prove that the pipeline is stronger than the demo.
Briefingshow
For AI agents, data access is often the boring bottleneck: they can calculate, explain and compare, but burn context on search, scraping and cleanup. FactIQ targets that layer by normalizing economics and finance data into an agent-friendly schema. If it works reliably, the coding agent becomes closer to a junior research analyst than a chat window with SQL access.
Sources