OYKH AI is an automated market-research service for Taiwan-listed equities. Before the open and after the close, it consolidates institutional (foreign / investment-trust / dealer) net flows, sector-level money rotation, active-ETF holdings changes, and a personal portfolio review into a single actionable brief — not one more dashboard of raw numbers.
Retail investors are not short of data — they are short of finished research. OYKH AI turns public market data into a daily conclusion with traceable numbers.
Previous-session institutional flows, foreign futures open interest, sector momentum, US market and FX context → today's watch list and risk notes.
Institutional net buy/sell, the ten focus sectors accelerating in or out, leading buys/sells inside each sector, and sectors where selling pressure eased.
Daily snapshots of active ETF holdings, compared across holdings, sector and trend layers to surface names and sectors multiple funds are adding together.
Your own watch list is reviewed every evening against that day's sector and index context — meaning, not just closing numbers.
Briefs are delivered automatically to Telegram / LINE, with the full archive kept on the web.
Each brief states its data date and comparison session. Figures come from the source data — never recomputed, never invented; unknown fields are left blank.
A fully automated research pipeline: ingest → align & compare → AI interpretation → delivery. No manual step.
Exchange and OTC market data, futures positioning, public disclosure filings and financial news are fetched and normalised daily.
Everything is framed as "today vs previous trading session" at sector and ticker level, so a single day's number is never read out of context.
Claude serves as the analysis core, turning structured data into causal, actionable conclusions — and flagging what is uncertain.
Scheduled jobs publish to Telegram / LINE and store the full archive for review, building a reproducible research trail.
Running daily in production for internal use and a small test group. Quality first, then more users.
Pre-open and post-close briefs, active-ETF snapshots and portfolio review all generate automatically with a stored history.
Personal watch lists, price and flow alerts, and a web dashboard so each user gets their own daily research.
Active-ETF consensus signals, sector rotation tracking, and institutional-flow data access — moving toward a subscription service.
We are looking for investors who will actually use the briefs daily and tell us what is wrong with them. One line about how you invest is enough.