# Your agent builds your CRM. We bring the data.

Connect your own Supabase as memory. Your agent pulls data through looot, stores it with its source and cost, and builds the interface.

## The short answer

Yes, an AI agent can build your own CRM. Give it a Supabase project for memory and looot for data. It creates the tables, fills them through endpoints billed per call, and builds a small dashboard on top.

- What it is. One key to the whole looot catalog.
- What it costs. Per call. The price is shown before a call that costs more than one cent.
- How it runs. Your agent calls looot and writes to your Supabase.

## How it fits

Four blocks: the agent, looot, your Supabase and your app. Your data stays in your project.

## The tables

- accounts: id, name, domain, industry, size, source, endpoint, cost_usd, fetched_at
- contacts: id, account_id, full_name, title, email, source, endpoint, cost_usd, fetched_at
- signals: id, account_id, kind, value, first_seen, source, endpoint, cost_usd, fetched_at
- runs: id, job, rows_added, rows_changed, source, endpoint, cost_usd, fetched_at
- notes: id, account_id, body, author, source, endpoint, cost_usd, fetched_at

## The prompt

Paste this into your agent. Swap the first line for any of: CRM, signal engine, ad tracker, lead list. Fill in your daily budget for [X].

You are building my CRM. Use looot for all data and my Supabase for memory.
1. Setup: set up https://looot.ai/skill.md, then connect my Supabase.
2. Create tables: accounts, contacts, signals, runs, notes. Every row stores source, looot endpoint, cost_usd, fetched_at.
3. Pipeline: find, enrich, verify, upsert. Upsert on a natural key. Never overwrite columns I edited.
4. Before any call that costs more than $0.01, show me the price and wait. Stop at my daily budget of $[X].
5. Watch loop: re-run each signal on a schedule, write only what changed, and tell me the diff.
6. UI: build a small dashboard over these tables (list, filters, row detail, run history).
7. Report each run: rows added, rows changed, total cost.

Step 5 needs scheduled runs, not available yet. Until then, ask your agent to run the loop when you want a fresh pass.

## Seven rules the prompt enforces

- Memory is your Supabase. Rows live in your project. looot keeps none of them.
- Every row has a source. Each row stores source, endpoint, cost_usd and fetched_at.
- Upsert on a natural key. Running the pipeline twice does not create duplicates.
- Your edits stay. The agent never overwrites a column you edited.
- Price before any call over one cent. It shows the price and waits for your yes.
- A daily budget. It stops at the amount you set.
- A report on every run. Rows added, rows changed and the total cost.

## What people build

- CRM. Accounts, contacts, notes, one list.
- Hiring signal tracker. Open roles per account, week over week.
- Ad swipe file. Saved ads with copy, date and source.
- Backlink monitor. New and lost links, flagged on each run.
- Creator database. Handles, followers, last post, contact.
- Market map. Companies in one space, grouped and sized.

## Can an AI agent build my own CRM?

Yes. Give it the setup file, a Supabase project and the build prompt on this page. It creates the tables, fills them through looot, and builds a small dashboard over them.

## Where is my data stored?

In your own Supabase project. looot returns data to your agent and keeps none of your rows.

## What does it cost?

You pay per call to looot. The prompt tells the agent to show the price and wait before any call that costs more than one cent, and to stop at your daily budget.

## Which tables does the prompt create?

accounts, contacts, signals, runs and notes. Every row stores its source, the looot endpoint, the cost and the fetch time.

## Can it re-run on a schedule?

Not yet. Step 5 of the prompt needs scheduled runs, which are not available. Ask your agent to run the loop when you want a fresh pass.

## What if I edit a row by hand?

The prompt says never to overwrite columns you edited, and to upsert on a natural key so a second run does not duplicate rows.

## Do I need to code?

No. Your agent writes the tables and the interface. You review what it built.

## What else can I build with it?

A hiring signal tracker, an ad swipe file, a backlink monitor, a creator database or a market map. Swap the first line of the prompt.