SIGLENS
Read the chart, not the indicators. AI analysis for US stocks, Korean stocks, and crypto.
Why SIGLENS
Reading a chart well takes more work than it looks. You add indicators one by one, check volume, look for patterns, skim the news, and then try to weigh all of it together. Most people stop halfway.
SIGLENS does that reading for you. It gathers price data, indicators, patterns, news, and market sentiment, and turns them into a report written in plain language. You can open a single symbol and read its analysis, or simply ask a question and get an answer.
SIGLENS never places an order. It offers analysis only, and every investment decision remains your own. You can start without an account; signing in adds a few extras described below.
Two ways to use it
siglens.io: analysis by symbol
Enter a ticker such as AAPL, 005930.KS, or BTCUSD. The page shows a chart with indicators, detected patterns, and an AI report that streams in as it is written. Tabs next to the chart cover news, the fear and greed index, and a combined outlook. Stocks add fundamentals and financial statements, and US stocks add the options market, congressional trading, and pre-market and after-market quotes. The AI report can be written by Claude Haiku, Sonnet or Opus 5.5, GPT-6 Luna, GPT-6 Sol or GPT-6.1 Sol, or Gemini 4 Argon.
Beyond single symbols, the site offers a watchlist, a sector signal dashboard, a market-wide fear and greed index, a news hub, a macroeconomic dashboard, and backtesting results for past AI analyses.
flowchart LR
A([Enter a ticker]) --> B[Identify the market]
B --> C[Chart and indicators]
C --> D[AI report]
D --> E[News, sentiment, and more tabs]
classDef step fill:#dbeafe,stroke:#2563eb,color:#1e3a8a
classDef start fill:#1e3a8a,stroke:#1e3a8a,color:#ffffff
class A start
class B,C,D,E step
ai.siglens.io: ask in your own words
Ask something like "How has Bitcoin been moving lately?" The AI agent decides what it needs, looks up the price, chart, existing SIGLENS analysis, and recent news, and then answers with the numbers it found. When no recent analysis exists, it runs a new one. Every step it takes is shown above the answer, along with how long it took.
Guests can ask a limited number of questions each day. Signing in removes that limit, saves your conversations, and lets the agent look at the holdings you registered.
flowchart LR
Q([Your question]) --> AG[AI agent]
subgraph LOOKUP[Looks up what it needs]
direction TB
P[Price]
CH[Chart and indicators]
N[News]
end
subgraph ANALYSIS[Finds an analysis]
direction TB
S[Saved SIGLENS analysis]
F[New analysis if none exists]
S -. not found .-> F
end
AG --> LOOKUP
AG --> ANALYSIS
LOOKUP --> R([Answer])
ANALYSIS --> R
classDef io fill:#4c1d95,stroke:#4c1d95,color:#ffffff
classDef agent fill:#ede9fe,stroke:#7c3aed,color:#4c1d95
classDef item fill:#ffffff,stroke:#94a3b8,color:#1e293b
class Q,R io
class AG agent
class P,CH,N,S,F item
style LOOKUP fill:#f1f5f9,stroke:#94a3b8,color:#1e293b
style ANALYSIS fill:#f5f3ff,stroke:#a78bfa,color:#1e293b
How accurate is it? We asked the agent and three general AI models (GPT, Gemini, Claude) the same 185 US and Korean stock questions and graded the answers against market and filing data. The results, every question, and every raw answer are on the benchmark page. The method is in BENCHMARK.md.
Signed-in members can also receive a regular email report on the symbols they follow. Guests can preview what it looks like on the sample page.
Supported markets
Each symbol belongs to one market, and the market decides the currency, trading hours, available timeframes, and which tabs appear.
| US stocks | Korean stocks | Crypto |
|---|
| Example | AAPL | 005930.KS, 247540.KQ | BTCUSD |
| Currency | USD | KRW | USD |
| Trading hours | NYSE and NASDAQ | KRX | Around the clock |
| Quotes | Real time | 20-minute delay | Real time |
| Fundamentals and statements | Yes | Yes | No |
| Options and congressional trading | Yes | No | No |
Version
The current version is shown in the badge at the top. Changes for each release are listed on the Releases page.
For developers
How it fits together
flowchart LR
U([Browser]) --> CF[Cloudflare]
CF --> APP
subgraph APP[SIGLENS app]
direction TB
WEB[siglens.io]
AI[ai.siglens.io]
end
APP --> CORE[siglens-core<br/>analysis domain]
APP --> DATA[Market data<br/>and news]
APP --> LLM[AI models]
APP --> STORE[Database and cache]
classDef edge fill:#1e293b,stroke:#1e293b,color:#ffffff
classDef dep fill:#ecfdf5,stroke:#10b981,color:#064e3b
classDef site fill:#ffffff,stroke:#3b82f6,color:#1e3a8a
class WEB,AI site
class U,CF edge
class CORE,DATA,LLM,STORE dep
style APP fill:#eff6ff,stroke:#3b82f6,color:#1e3a8a
Both sites are served by the same application, which chooses what to render from the host name. Indicator math, pattern detection, signal logic, and prompt building live in @y0ngha/siglens-core, a separate package that holds the analysis domain. Analysis techniques are written as Markdown files under skills/, so a new technique can be added without touching calculation code.
Running locally
git clone https://github.com/y0ngha/siglens.git
cd siglens
yarn install
cp .env.example .env.local
yarn dev
The app starts at http://localhost:4200. Use the Node.js version in .nvmrc; the Yarn version is pinned in package.json. Installing dependencies requires a GitHub token because @y0ngha/siglens-core is published to GitHub Packages. Every environment variable is explained in .env.example, and every script is listed in package.json.
Documentation
Project documentation is written in Korean. The public documents live in docs/product/ and docs/conventions/:
Security
Please do not report a vulnerability in a public issue. Follow the process in SECURITY.md.
Contributing
The project does not accept outside code contributions yet. Bug reports and suggestions are welcome in Issues.
License
MIT License
