Downloads
Start with a .uwx.md record, the complete structured underwriting file. No account, connector, or preferred editor is required: compatible agents, services, editors, and viewers can all load it. See UWX and UW Lite for the compact .uw.md summary representation.
Create a blank record
The quickest route is the published CLI, which fills in the identifying fields for you:
npx @uwmd/cli init --output my-deal.uwx.md --asset-class office --tier analystSee the quickstart for the next steps.
Blank templates
If you would rather start from a file, both templates are Format 3.0 UWX records generated by uwmd init. They validate as they are, with one warning asking for the property's unit count.
- Download the blank screener template (
.uwx.md) — a first-pass screening record. - Download the blank analyst template (
.uwx.md) — the same structure marked for full underwriting.
Both default to multifamily. Replace the template deal_id and dates, and change asset_class when appropriate.
Programs
- npm packages. Published in release 4.3.0:
@uwmd/verify,@uwmd/core,@uwmd/cli,@uwmd/signing,@uwmd/batch,@uwmd/excel,@uwmd/report,@uwmd/lake. See tools and packages for what each one does. - Download the browser viewer (
.html) — save it, open it in a browser, and drop a.uwx.mdrecord onto the page. - Download all source code (
.zip) — includes every package, the web tools and VS Code extension, plus schemas and tests. - GitHub releases — tagged release notes.
Example deals
- Multifamily
- Multifamily, UW Lite summary (
.uw.md) - Office
- Retail
- Industrial
- Self-storage
- Hospitality
- Senior housing
- Student housing
- Land
- Data center
- Manufactured housing
- Mixed-use (RFC 0019 components)
- Capital stack (RFC 0026 senior + mezz + pref)
AI instruction files
Add the relevant file to an AI project or paste its contents into the platform's project instructions.
- Codex / agent skill (
SKILL.md) - Claude project instructions (
CLAUDE.md) - ChatGPT project instructions (
.txt) - Gemini project instructions (
GEMINI.md) - Platform-neutral AI guide (
.md)
For automated discovery, use llms.txt, the expanded llms-full.txt, or the AI information page.