UW Markdown for AI and code

UW Markdown is an open interoperability standard, not a hosted AI service or a single application. AI agents, deterministic calculation services, internal underwriting systems, and compatible editors or viewers can load the same deal record directly. No custom connector or MCP server is required.

Core rules

1. Preserve the Markdown narrative and labeled JSON blocks.
2. AI may extract source facts, classify information, identify gaps, and draft narrative.
3. AI must not calculate financial results. Use deterministic code or formulas for NOI, DSCR, LTV, debt yield, cap rate, IRR, NPV, and DCF.
4. Preserve bytes outside the requested edit region for Tier-2 edits.
5. Preserve provenance. Append or supersede; do not silently destroy history.
6. Rates are fractions (0.055 means 5.5%).
7. Treat the format spec, protocol, schemas, and conformance fixtures as the contract.

Best context order

1. llms.txt for discovery.
2. First-file tutorial for orientation.
3. Format specification for document syntax and semantics.
4. Protocol specification for reader, editor, calc-host, and agent-host behavior.
5. Schemas and conformance fixtures for implementation testing.
6. llms-full.txt for the expanded link map.

Optional HTTP and MCP profiles

The HTTP binding, OpenAPI contract, and MCP binding describe interoperable server shapes. They are optional profiles. This site does not expose live deal resources or mutation tools.

The MCP profile defines uwmd.get_document, uwmd.validate, uwmd.convert, uwmd.apply_edit, and uwmd.list_representations.

Ready-to-use instruction files

• Codex-compatible skill
• Claude instructions
• ChatGPT project instructions
• Gemini instructions
• Platform-neutral guide

Canonical project

• Repository: https://github.com/UWMD-OSP/UW-Markdown
• License: MIT
• Authoring: UW Format 2.0 (.uwx.md for structured records).
• Compatibility: see the current version matrix for protocol and package versions.