Feed the Vault documents
Updated July 11, 2026
Rather than typing every appetite row, fee line, standard, and vendor into the Vault by hand, you can drop a document on the Vault page and let the engine pull the structure out of it for you.
#Drop a file
The Documents section of the Vault carries the instructions right on the page: Drop a markdown, TXT, PDF, or CSV anywhere on this page and the engine pulls the appetite rows, fees, standards, and vendors out of it for your review. Matching entries get replaced, new ones get added. The same note points you at a sample file to try it with: Try it: download the sample SOP file and drop it back in.
Markdown, plain text, and PDF files are all read for appetite rows, fee lines, standards, and vendor entries. A CSV file is read only for appetite rows, columns mapped straight in; fees, standards, and vendors stay untouched by a CSV drop.
This is an owner-only action, the same as the rest of Vault editing. See Edit the Vault for the full picture of who can change what.
#The review modal
After the engine reads your file, it does not merge anything automatically. You land on a review screen first, headed Here is what I found. Merge it in?, with an intro line naming the source document: Pulled from <file name>. Check what goes in. Matching entries get replaced, new ones get added.
Findings are grouped by section, each with a count: appetite rows found, fee lines found, standards found, vendors found. Every row inside those groups carries one of two tags:
- replaces current, when the row matches something already in your Vault
- new, when nothing on file matches it yet
Every checkbox on this screen starts checked. Uncheck anything you do not want pulled in, then click Merge into the vault to commit, or Cancel to walk away and leave your Vault exactly as it was.
#What merging actually does
For everything still checked when you confirm: a matching appetite row, fee line, or vendor gets overwritten in place with the new version, and anything with no match gets added as a new row. A standard that is not already present word for word gets added; standards are never overwritten, only added if new. The source document itself gets recorded in the Documents list, so you can see later where a given entry came from.
#When nothing vault-shaped is found
If the file you dropped has nothing the engine recognizes as an appetite row, a fee line, a standard, or a vendor, you get a plain message instead of a review screen: Nothing vault-shaped found in <file name>. Appetite tables, fee lines, standards, and vendor lists are what it reads.