FactIQ

Financial research questions rarely stay in one data domain. FactIQ is a hosted data plugin — accessed via the Claude Code plugin marketplace or Codex — that lets an AI agent pull company fundamentals from SEC filings, live and historical market prices, commodities and FX rates, and official economic time series from global institutions, all inside a single session. The vendor demonstrates this with a Micron thesis that decomposes into quarterly revenue from 10-K filings, DRAM and NAND export volumes from Korea Customs by HS code, and flash and DRAM data from Japan Customs — one question, three jurisdictions, every number cited to source.

The differentiating architecture is source traceability. Rather than returning a synthesized answer, the plugin structures its output so each data point links back to the filing, customs bureau, or institution it came from. For anyone stress-testing an investment thesis or building an agent that downstream users need to audit, that trace matters more than the number itself.

The free tier covers the full current data catalog — approximately 1,000 US-listed companies above $10B market cap, market prices, and the 25M-plus macro time series — with no stated limits on queries, charts, or seats. The paid tier is on a waitlist and gates access to bespoke disclosures extracted from 10-Q and 10-K filings, structured earnings call and investor event insights, government policy document intelligence, satellite imagery, and tabular data from industry PDFs. Teams that need those layers before the waitlist resolves will need to source them separately.

Setup follows three steps the docs describe: add the plugin in the terminal, run the reload and MCP commands inside Claude Code, and complete a browser-based account sign-in — no API key required. There is no self-hosted option; the data service runs on FactIQ’s infrastructure, which means data freshness, uptime, and any future pricing changes are outside the user’s control.