How to build a customer feedback board with Claude

Set up a customer feedback board for me in databasedmcp, in one go.

1. Create a database called feedback, described as "Customer feedback board: feature requests, the customers who asked and what they said, filed by Claude."
2. Create three tables:
   - requests: id, title (required), detail, status (under_review, planned, in_progress, shipped or declined, defaults to under_review), created_at (defaults to now), shipped_at.
   - customers: id, name (required), company, plan.
   - votes: id, request_id (references requests, required), customer_id (references customers, required), source (call, email, support or chat), quote (their words), at (defaults to now). One vote per customer per request: make request_id and customer_id unique together.
3. Write the database a guide that says what each table is for, to search request titles and details before adding a new request because most feedback matches one already there, to look customers up by name and company before adding them, that a customer votes once per request, and the queries for "most wanted open requests" and "who to tell when a request ships".

From then on: when I paste feedback, match it to an existing request or add a new one, add the customer if they're new, and log their vote with the source and a quote in their own words. If they ask again, replace their quote instead of counting a second vote. When I say a request's status changed, update it. Show me the tables when you're finished.

CREATE TABLE requests (
  id INTEGER PRIMARY KEY,
  title TEXT NOT NULL,
  detail TEXT,
  status TEXT NOT NULL DEFAULT 'under_review' CHECK (status IN ('under_review', 'planned', 'in_progress', 'shipped', 'declined')),
  created_at TEXT NOT NULL DEFAULT CURRENT_TIMESTAMP,
  shipped_at TEXT
);

CREATE TABLE customers (
  id INTEGER PRIMARY KEY,
  name TEXT NOT NULL,
  company TEXT,
  plan TEXT
);

CREATE TABLE votes (
  id INTEGER PRIMARY KEY,
  request_id INTEGER NOT NULL REFERENCES requests(id),
  customer_id INTEGER NOT NULL REFERENCES customers(id),
  source TEXT CHECK (source IN ('call', 'email', 'support', 'chat')),
  quote TEXT,
  at TEXT NOT NULL DEFAULT CURRENT_TIMESTAMP,
  UNIQUE (request_id, customer_id)
);

# feedback: requests, customers, votes

- **requests**: one row per feature request. Search titles and details (LIKE) before adding a new one; most feedback matches something already here. status: under_review, planned, in_progress, shipped, declined. Set `shipped_at` when it ships.
- **customers**: one row per customer who asked for something. Look up by name and company first.
- **votes**: one row per customer per request, with `quote` in their words and `source` (call, email, support, chat). A customer votes once per request.

Most wanted: `SELECT r.id, r.title, r.status, count(v.id) AS votes FROM requests r LEFT JOIN votes v ON v.request_id = r.id WHERE r.status NOT IN ('shipped', 'declined') GROUP BY r.id ORDER BY votes DESC`

Who to tell when a request ships: join votes to customers on `request_id`.

INSERT INTO customers (name, company, plan) VALUES (?, ?, ?) RETURNING id
-- params: ["Priya Raman","Kestrel Analytics","Team"]

INSERT INTO requests (title, detail) VALUES (?, ?) RETURNING id
-- params: ["Export reports as CSV","Download any report as a CSV file to use in spreadsheets."]

INSERT INTO votes (request_id, customer_id, source, quote) VALUES (?, ?, ?, ?) ON CONFLICT (request_id, customer_id) DO UPDATE SET quote = excluded.quote
-- params: [1,1,"email","Our finance team rebuilds every report by hand in a spreadsheet each month."]

UPDATE requests SET status = 'planned' WHERE id = ?
-- params: [1]

SELECT r.id, r.title, r.status, count(v.id) AS votes FROM requests r LEFT JOIN votes v ON v.request_id = r.id WHERE r.status NOT IN ('shipped', 'declined') GROUP BY r.id ORDER BY votes DESC, r.created_at LIMIT 10

SELECT c.name, c.company, c.plan, v.source, v.quote, v.at FROM votes v JOIN customers c ON c.id = v.customer_id JOIN requests r ON r.id = v.request_id WHERE r.title LIKE ? ORDER BY v.at
-- params: ["%CSV%"]

SELECT r.title, count(*) AS votes FROM votes v JOIN customers c ON c.id = v.customer_id JOIN requests r ON r.id = v.request_id WHERE c.plan IN ('Team', 'Business') GROUP BY r.id ORDER BY votes DESC