Please build out my FlipperForce Material Catalog from the purchase history in [FILE NAME].
1. Read the file and list what you found: how many rows, what columns are present, and the date range covered. Create nothing yet.
2. Pull my existing Material Catalog so you know what is already in there.
3. Work out which rows are real catalog material. Drop consumables, tool rentals, delivery fees and one-off purchases I would never reuse, and tell me what you dropped and why.
4. Collapse repeat purchases of the same item into one material. Use the most recent price, and tell me where an item was bought at more than one price.
5. For each material, map the fields: title, price per unit, unit of measure, category, SKU, product URL, thumbnail and supplier company. Use only the standard FlipperForce material categories. If a row has no clear category, use misc rather than guessing.
6. Show me the full list as a table before you create anything: title, price, unit, category, supplier. Flag any row where you had to infer the price or the unit.
7. Wait for my confirmation. Create only the rows I approve.
8. Skip anything that already exists in my catalog rather than creating a duplicate, and list what you skipped.
9. When you are done, confirm how many materials were created and list anything that failed.
Do not change or delete anything already in my catalog.
There is no bulk upload endpoint. Each material is created individually, so a large import is a lot of requests and can hit rate limiting. If a run stops partway, the materials already created stay put and your assistant can resume from where it left off.
Categories are a fixed FlipperForce list. Your assistant maps to the closest match and uses misc where nothing fits, rather than inventing a category that does not exist.
The supplier company field links to your Companies list, so add the supplier as a company first if you want purchases attributed to it.
The Material Catalog only pays off once it has your real products in it at your real prices. Getting there means finding each item on a supplier site, copying the link, pasting it in, and repeating that a few hundred times.
So most catalogs stall at twenty items. The estimator keeps falling back to national average allowances and your material numbers stay generic, when you already have years of actual purchases sitting in your supplier account.
Your assistant reads your supplier export, works out which rows are reusable material rather than consumables and fees, collapses repeat purchases of the same product into a single item at its most recent price, and maps each one to the fields the Material Catalog expects.
You see the full list as a table first. Once you approve it, the materials are created in your catalog with their prices, SKUs, product links and thumbnails attached, so the estimator can pull your real numbers instead of national average allowances.
What it will not do: it does not change or delete anything already in your catalog, does not create a duplicate of a material you already have, does not invent a price or a unit it could not read off the export, and does not create anything before you have seen the list.
This prompt requires a connection to the FlipperForce Public API. Once your AI agent is connected to your account, you can run any prompt in the library.
Learn how to connect your AI agent to the FlipperForce Public API
Before you run this, check:
Run it on one supplier at a time. A single Home Depot export is a clean first pass, and you can see what the mapping looks like before pointing it at three years of everything.