nashnova: An AI Investment Research Workbench for A-Share, Hong Kong, and US Markets

What is nashnova

nashnova is an AI-powered investment research workbench built for individual investors, covering the A-share, Hong Kong, and US markets. It ships with more than twenty built-in strategies and nearly a hundred expert personas, including the "Fu Peng Persona" co-released with macroeconomist Fu Peng. You describe a research task in plain language, and the system automatically breaks it into steps, pulls the data, runs the calculations, and produces the deliverable — and the task keeps running even after you close the page. It supports strategy backtesting, report generation, supply-chain breakdowns, and continuous portfolio tracking. Key figures are always traceable to their source, and results can be exported as PDF or shared via a link.

Key features

• Built-in financial database: Preloaded A-share, Hong Kong, and US market data spanning stocks, ETFs, and commodities — no data source to wire up yourself. It covers quotes and technical indicators, financial statement triples, earnings-call transcripts, announcement events, broker research, and financial news, with every key number tagged by source and timestamp for line-by-line verification.
• Expert personas: Nearly a hundred personas mapped to different investment methodologies. For complex questions you can summon macro, industry, and value perspectives at once and compare the variables each one watches.
• Strategy backtesting: More than twenty built-in strategies across technical indicators, momentum, mean reversion, grid, and event-driven approaches, with backtesting for stocks, ETFs, and commodities. The system handles data pulls, date alignment, grouped calculations, and chart generation automatically.
• Research report generation: Four templates — company one-pager, supply-chain map, earnings review, and portfolio check-up — that auto-source evidence and produce charted reports you can export as PDF or share via link.
• Market-pattern validation and supply-chain breakdown: Test hypotheses like "is there abnormal return before an industry's earnings disclosure" on historical data, with significance and methodology notes; or map an industry's upstream/downstream links, representative firms, and value distribution into a shareable graph.
• Scheduled tasks and continuous tracking: Run daily, on weekdays, weekly, or once, each with its own timezone. Stays silent when nothing material changes, alerting only when holdings or earnings actually move.
• Long-term memory: Holdings, watchlists, preferences, and past research live in your workspace and carry across sessions, so you never re-explain context.
• Custom research methods: Save your recurring research steps and judgment criteria as a personal capability you can reuse on similar tasks.

How to use nashnova

1. Sign up: Visit https://www.nashnova.com/ or download the iOS/Android app, create an account, and enter the research workbench.
2. Describe a research task: Type your need in natural language; the system breaks it into steps and runs them, continuing even after you leave the page.
3. Invoke expert personas: For complex questions, summon macro, industry, and value personas together and compare their key variables.
4. Run a backtest: Pick from twenty-plus built-in strategies; the system pulls data, aligns dates, groups the calculations, and charts the results.
5. Generate a report: Use one of the four templates to auto-source a charted report, exportable as PDF or a share link.
6. Verify sources: Every key figure carries source and timestamp — click through line by line to keep conclusions traceable.
7. Set scheduled tasks: Configure daily, weekday, weekly, or one-off runs that alert only on material change.
8. Save personal methods: Persist your usual steps and criteria as a custom method for reuse.
9. Resume across sessions: Holdings, watchlists, and preferences auto-save to long-term memory, so weeks later "how did that strategy do" picks up right where you left off.

Core advantages

• Full-market coverage: One stop for A-share, Hong Kong, and US markets across stocks, ETFs, and commodities — no data plumbing required.
• Autonomous task execution: Describe a task and the system decomposes, fetches, computes, and delivers, running on after you close the tab.
• Traceable data: Every key figure is sourced and timestamped, so results are reproducible and reusable.
• Expert persona roster: Talk macro and asset allocation with the "Fu Peng Persona" from a well-known economist.
• Multi-perspective cross-checking: Call different methodologies on the same question to cut single-viewpoint bias.
• Backtesting out of the box: Twenty-plus strategies with automatic data pulls, date alignment, and grouped runs.
• Alerts only on change: Scheduled tasks stay quiet unless something material moves, avoiding notification noise.

Comparison with Tonghuashun iWencai

Use cases

• Batch screening: Turn screening or timing rules into executable logic, scan the whole market on schedule, and output candidate lists with evidence.
• Validate hypotheses: Backtest rules like "buy after an earnings beat and hold" for significance, applicable range, and failure conditions.
• Break down supply chains: Quickly map upstream/downstream, representative firms, and value split when entering a new industry, into a shareable graph.
• Multi-perspective cross-check: Call different methodologies on the same question and compare key variables.
• Portfolio and watchlist tracking: Schedule tracking that alerts only on material change.

Caveats

nashnova provides market intelligence and research support, not personalized investment advice. Conclusions rest on historical and public data; investing carries risk and decisions are yours to make.

FAQ

Who is nashnova for? Individual investors who want to systematize their research workflow, test their own ideas, and keep tracking holdings — especially those needing unified research across A-share, Hong Kong, and US markets.

How is nashnova different from ordinary stock-screening tools? It's not just screening Q&A. It decomposes a research task into executable steps, auto-fetches and backtests, generates evidence-backed reports, and supports long-term memory and cross-session continuity — closer to a runnable research system.

Is the data source reliable? The built-in financial database covers quotes, financials, transcripts, announcements, research, and news, with every key figure tagged by source and timestamp for line-by-line checks, so results are traceable and reusable.