We combine deep supply chain intelligence with AI-powered analytics to identify asymmetric opportunities across Chinese and global markets.
我们将深度供应链情报与 AI 驱动的分析相结合，在中国和全球市场中识别不对称投资机会。
KSINQ operates at the intersection of cross-border trade and multi-asset investment. With operational infrastructure spanning mainland China, Hong Kong, and the United States, we bring a practitioner's perspective to global markets.
KSINQ 在跨境贸易与多资产投资的交汇点运营。我们的运营基础设施横跨中国大陆、香港和美国，以实操者的视角审视全球市场。
Our edge is structural: real trade flows inform our macro view, AI-native tooling accelerates our research, and multi-jurisdiction access enables direct market participation where others rely on intermediaries.
我们的优势是结构性的：真实贸易流为宏观观点提供信息，AI 原生工具加速研究，多法域准入使我们能够直接参与市场。
Explore our capabilities →探索我们的能力 →KSINQ is an AI-powered research platform that automates cross-border investment analysis. We ingest real-time market data, process it through customized LLM workflows, and deliver actionable research across multiple asset classes and jurisdictions.
KSINQ 是一个 AI 驱动的研究平台，自动化跨境投资分析。我们摄取实时市场数据，通过定制化的 LLM 工作流处理，为多个资产类别和法域提供可执行的研究成果。
Real-time feeds from AKShare (A-shares, ETFs), S&P Global/Kensho (global fundamentals), and 10+ financial data APIs covering equities, commodities, FX, and crypto.
来自 AKShare（A 股、ETF）、S&P Global/Kensho（全球基本面）及 10+ 金融数据 API 的实时数据流，覆盖股票、商品、外汇和加密货币。
Multi-model LLM workflows (Claude, GPT-4o, open-source models) orchestrated through Dify for fundamental analysis, earnings processing, and macro synthesis. Each model is routed to tasks matching its strengths.
多模型 LLM 工作流（Claude、GPT-4o、开源模型）通过 Dify 编排，执行基本面分析、财报处理和宏观综合研究。每个模型被路由到最适合其能力的任务。
Structured research output with probability estimates, Kelly-criterion position sizing, explicit falsification criteria, and stop-loss conditions. Every thesis is stress-tested through buy-side, risk, and trading perspectives.
结构化研究输出，包含概率估计、Kelly 准则仓位管理、明确的证伪条件和止损条件。每个投资论点都经过买方研究、风险管理和交易三重视角的压力测试。
Automated monitoring with trigger-based alerts for QDII/LOF premium shifts, macro regime changes, and cross-market arbitrage windows. Research delivered through API endpoints and structured dashboards.
基于触发器的自动化监控预警——QDII/LOF 溢价变动、宏观政策转向、跨市场套利窗口。研究成果通过 API 端点和结构化仪表盘交付。
Our infrastructure runs on Cloudflare (edge delivery & DNS), AWS (Bedrock AI inference & data pipelines), and Google Cloud (compute & analytics), with self-hosted services for latency-sensitive operations.
基础设施运行于 Cloudflare（边缘交付和 DNS）、AWS（Bedrock AI 推理和数据管道）及 Google Cloud（计算和分析），同时保留自建服务处理延迟敏感型操作。
From raw market data to actionable signals, we build the full stack of cross-border investment intelligence.
从原始市场数据到可执行信号，我们构建跨境投资智能的完整技术栈。
Custom LLM pipelines for fundamental analysis, earnings transcript processing, and multi-source synthesis across Chinese and global markets.
Learn more →Real-time and historical data across A-shares, HK stocks, US equities, futures, ETFs, and crypto-adjacent equities through custom API orchestration.
Learn more →Leveraging real trade flows and supply chain signals to inform macro and sector-level investment decisions. Ground truth from the factory floor.
Learn more →Systematic tracking of cross-listed fund premiums and arbitrage windows between onshore and offshore markets with automated threshold alerts.
Learn more →Every thesis requires probability estimates, odds assessment, position sizing, stop-loss conditions, and explicit falsification criteria.
Learn more →Tracking energy cycles, USD/UST dynamics, China macro, Middle East geopolitics, and semiconductor supply chains as structural drivers.
Learn more →Operating across jurisdictions with direct market access and regulatory infrastructure in mainland China, Hong Kong, and the United States.
横跨多个法域运营，在中国大陆、香港和美国拥有直接市场准入和监管基础设施。
Every investment thesis is stress-tested through three independent perspectives before capital is deployed.
每个投资论点在资金部署前，都经过三个独立视角的压力测试。
Fundamental analysis with causal reasoning. Separate facts from assumptions. Identify catalysts and build conviction through evidence, not narrative.
Tail risk assessment, correlation analysis, and position sizing. Every position carries a stop-loss and explicit criteria for what would prove the thesis wrong.
Liquidity conditions, execution timing, and market microstructure. Probability-weighted odds with asymmetric payoff profiles. Entry matters as much as thesis.
Our research stack combines large language models with real-time market data and proprietary analytical workflows.
我们的研究技术栈将大语言模型与实时市场数据及专有分析工作流相结合。
We welcome conversations about markets, technology, and cross-border opportunities.
我们欢迎关于市场、技术和跨境机遇的对话。
[email protected]