Strategy and AI workflows

What this workflow solves

It standardises how data is prepared for strategy research and AI workflows, which is where inconsistent feature definitions usually creep in.

An agent should read structured data, lineage and data_gaps first, rather than relying on a single natural-language summary.

Which APIs it uses

• features
• factor_data
• time_alignment

The minimal flow

• Establish the timeline and trading-day baseline from price data
• Add monthly revenue and reported financials for fundamental context
• Bring in institutional-flow and other trading-side data
• Keep data_gaps, confidence and source attribution for the agent to read

The order that works

Interpret, then model, then align. That order reduces the drift between a backtest and live execution.

Entry points an agent can read

For an agent reading the documentation and the data surface programmatically, /llms.txt, /llms-full.txt and /openapi.json are the entry points.

The MCP server is live at https://mcp.twmarketdata.com/mcp. It is a data access layer — it is not a trading system and it does not issue investment advice.

Scope of this workflow

• This workflow is for data research, strategy analysis and systems integration. It is not investment advice.
• Avoid emitting buy/sell/hold instructions or target prices.