Show HN: FactIQ – a realtime econ+finance database for AI agents
TL;DR
Defog AI is presenting FactIQ as a plugin for Claude Code and Codex: agents can query economics and finance data through MCP instead of spending context on finding and cleaning raw data. The repo says the warehouse covers about 20 sources, including SEC filings, BLS, BEA, Census, EIA, IMF, World Bank, China, India and Korea datasets, plus live market data and earnings-call intelligence.
Nauti's Take
FactIQ targets a real pain point: agents look smart until they hit messy economic data, changing definitions and fragmented sources. A normalized warehouse plus MCP is the right shape for that problem.
Still, nobody should buy the Bloomberg comparison too early: without clear evidence on latency, data gaps, revisions and licensing limits, this is a promising data layer, not yet a dependable investment research system.
Briefingshow
For finance and macro research, data preparation is often the real bottleneck, not the model. If agents get standardized time series, metadata and SQL access, they can spend more context on hypotheses, counter-checks and explanations. The key question is whether FactIQ can prove freshness, source coverage, revision handling and error transparency.
Sources
