Skip to main content
Each project can have a SQLite database hosted on Cloudflare D1. Store user profiles, conversation history, leads, or any data your WhatsApp application needs.

​
Enable database

  1. Go to project settings → Database
  2. Click “Enable database”
  3. Database provisions automatically via Cloudflare D1

​
Dashboard features

  • Spreadsheet view - Click cells to select, double-click to edit
  • Table management - Create tables, add/rename/drop columns
  • SQL editor - Run raw queries
  • CSV import/export - Bulk data operations
  • Pagination - 25/50/100/200 rows per page

​
Column types

TypeDescription
textString values
integerWhole numbers
numberDecimal numbers
booleantrue/false
dateDate without time
datetimeDate with time
jsonJSON objects
arrayJSON arrays
file_urlURL to file
referenceForeign key

​
Using the database

​
Platform API

Query and modify data from your backend via REST endpoints. See Database API. Agent steps can access the database by adding a webhook tool that calls these endpoints.

​
Functions

Cloudflare Worker functions can access the same project database directly via env.DB: 
async function handler(request, env) {
  const body = await request.json().catch(() => ({}));
  const vars = body.execution_context?.vars || {};

  const { results } = await env.DB.prepare(
    "SELECT * FROM customers WHERE phone = ?"
  ).bind(vars.phone).all();

  await env.DB.prepare(
    "INSERT INTO audit_logs (event_type, payload) VALUES (?, ?)"
  ).bind("customer_lookup", JSON.stringify({ phone: vars.phone })).run();

  return new Response(JSON.stringify({
    customer: results[0] || null
  }), {
    headers: { "Content-Type": "application/json" }
  });
}
If the function runs as an agent tool, the agent-controlled arguments are inside body.input. If it runs as a workflow function, use body.execution_context, body.flow_info, and body.flow_events. Agent steps can invoke functions as tools to access the database indirectly.