Get started with Wabery
Este conteúdo não está disponível em sua língua ainda.
First choose the customer outcome and who will answer. Then test that setup in a private Wabery chat or a real WhatsApp Sandbox conversation. You do not need a Meta account or business number for either test.
Create a free account1. Create your project
Section titled “1. Create your project”Sign up at app.wabery.com. Wabery creates your first project during onboarding. You do not need a credit card to create the account; paid delivery and AI usage have their own funding terms.
One project normally belongs to one client or product. Its contacts, conversations, channels, and credentials stay together.
2. Choose what customers should be able to do
Section titled “2. Choose what customers should be able to do”For a Wabery hosted agent, choose Customer support, Lead qualification, Reservations, Appointment requests, Online store, or Custom. You can add or turn off Leads and Bookings later without rebuilding the project.
Then choose who writes replies.
Wabery AI agent
Section titled “Wabery AI agent”Choose the Wabery AI agent when you do not already have a deployed chatbot. Wabery asks what it should do, how it should sound, which customer language it should use, and what it should know. It creates a Draft you can test before anything reaches a live client channel.
Wabery runs this agent for you. Hosted AI usage comes from your Wabery wallet. Follow the hosted-agent setup →
Your own backend or external agent
Section titled “Your own backend or external agent”Choose I already have a backend when n8n, Make, Zapier, your application, or another AI agent already writes the replies. Wabery gives that system the WhatsApp connection:
WhatsApp → signed Wabery webhook → your system → Wabery send API
You will pick the backend type, add its public webhook URL, verify a test event, and use a reply-ready API key. Your system must stay online to answer. Follow the external-agent setup →
3. Finish the setup
Section titled “3. Finish the setup”| You chose | Do this next |
|---|---|
| Wabery AI agent | Follow Build a Wabery hosted agent. Keep it as a Draft for the sandbox test. |
| Your own backend | Follow Connect your external agent. Verify the webhook and turn on Use my API. |
4. Try the sandbox or a private test
Section titled “4. Try the sandbox or a private test”For a hosted agent, start with the embedded simulator or create a private Wabery link. A private link sends no WhatsApp message. You control its expiry and message limit, and you can disable it at any time.
Use the shared WhatsApp Sandbox when you want to test the full phone experience.
Open Channels → WhatsApp Business → Sandbox, copy the join code (it looks
like wab-ABC123), and send it from your own phone.
Then send a normal message such as hello. That creates the first conversation
in Inbox. The join code itself is not a chat bubble.
The sandbox tests the reply mode you just configured. Full detail: Wabery sandbox.
5. Check the reply
Section titled “5. Check the reply”Send a customer-style message. Your hosted agent or external backend should answer it. Open Inbox to see the conversation or take over manually.
The Quickstart shows how to run this setup with an AI coding agent or through the website.
6. Go live when you are ready
Section titled “6. Go live when you are ready”When the tests work, open Channels and connect the destinations the project needs. This can be the client’s WhatsApp Business number, Website chat, or a supported beta social channel. Each connection remains separate from agent publication.
For hosted agents: edit the Draft, test it, then review and publish on Agent. External tools cannot publish Live. Turn on Leads separately, and publish booking rules before launching agent booking actions when the project uses Reservations.