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.

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

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

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.

The data tools behind the AI agent are also available over HTTP. Create an API key in your account settings, then call them from a script or hand them to your own agent.

• API reference: siglens.io/developers covers authentication, rate limits, errors, and every tool with a curl example.
• OpenAPI 3.1: openapi.json and openapi.yaml, generated from the same tool list as the API.
• Agent Skill: SKILL.md lets Claude Code or Claude.ai call the API on its own.

AI analysis runs on your own LLM key (BYOK), so SIGLENS does not pay for it, and reading an analysis that is already cached costs nothing. Every model can generate a new analysis, but only on your own key for that model's provider: register a key first, and SIGLENS's own LLM key is never used. The website generates the lower models with its own key; the API does not.

Each symbol belongs to one market, and the market decides the currency, trading hours, available timeframes, and which tabs appear.

The current version is shown in the badge at the top. Changes for each release are listed on the Releases page.

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.

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.

Project documentation is written in Korean. The public documents live in docs/product/ and docs/conventions/:

Please do not report a vulnerability in a public issue. Follow the process in SECURITY.md.

The project does not accept outside code contributions yet. Bug reports and suggestions are welcome in Issues.

MIT License