Skip to content

User Guide

This guide is for normal users - people who want to fetch data, compute indicators, build collections, run research, and share dashboards. You do not need to write code or modify the project.

What DAAS gives you

DAAS is a local data platform. Out of the box it ships with:

  • ~5,600 entities - 5,575 stocks + 60 countries, each linkable to its datasource identifiers.
  • Hundreds of prebuilt indicator rules (and the platform scales to thousands as you add datasources) - moving averages, RSI, rolling volatility, pct change, z-scores, and macro/economic series.
  • 8 MCP tool groups (alerts, cron, composite, daas, dashboard, leader, pdf, research) for browsing, scheduling, alerting, and orchestration.
  • 40+ skills that wrap common workflows (fetch, research, dashboard, collections, PDF search, web scraping).

How you talk to DAAS

You have two first-class paths; pick whichever fits the moment:

  1. Skills (recommended for first-time users) - in Claude Code, just describe what you want ("fetch AAPL daily prices and add a 20-day moving average"). The right skill resolves the entity, calls the Python library, and persists to daas.db. See First Fetch in 5 Minutes.
  2. MCP server - if your client speaks MCP, point it at fd-daas-mcp (the sole .mcp.json entry) and call the <group>_<tool> tools directly. See MCP Tools.

Both paths share one daas.db, so work done through one is immediately visible to the other.

  1. Install & Deploy - get the project running.
  2. First Fetch in 5 Minutes - your first real fetch + indicator.
  3. Examples - the seven canonical workflows.
  4. Concepts - understand entities, collections, indicators.
  5. Skills - the official skill families.

Concepts you will meet

  • Entity - a stock or country (e.g. AAPL, CN). See Entities.
  • Datasource - a source of data (akshare, yfinance, edgar, ...).
  • Indicator rule - a binding that computes a series (e.g. SPY_ma5 = 5-day SMA of SPY closes) into the observations table.
  • Collection - a named group of entities or indicators. See Collections & Indicators.
  • Research - a persisted bundle tying an entity collection, an indicator collection, a dashboard, and a pipeline to a markdown report.

No coding required

If you can describe what you want in plain language, a skill can do it. The examples use real commands so you can copy-paste when you prefer.