Whatsplaid
Language and currency
Start free
Plans
Search the site
Language and currency
External connector via API

Connect your system's data to the AI agent on WhatsApp.

Allow the agent to consult orders, products, registrations, and settings during the conversation. Your API provides the data; Whatsplaid uses the relevant information to help the customer.

Knowledge source API
Assistant on WhatsApp
From question to answer
01
The customer asks a questionThe agent detects when system data is required.
02
The agent collects the identifierEmail, mobile number or code, according to your settings.
03
Your API returns the dataThe connected system determines which information can be accessed.
04
The conversation continues with contextThe agent uses relevant data to formulate the response.

A knowledge source with the data your operation needs.

The connector can receive JSON objects or lists with different structures. You decide what to expose in the API and in which situations the agent should consult it, without limiting the integration to a single type of information.

Sales

Products and availability

Provide features, pricing, and availability to support questions about your catalog.

After-sales service

Orders and deliveries

Return the order status, items and delivery information authorized for the customer.

Relationship

Registration data

Make relevant customer registration information available to contextualize support.

Support

Account settings

Help the agent understand the customer's context with configuration, consumption, or limit data returned by your software.

Configure the connection and set when it joins the conversation.

The technical configuration is located under Developers. After testing and activating the connector, choose the API as the knowledge source in the AI agent settings.

01

Connect your API

In Developers, open External connectors and select Knowledge base via API. Enter the HTTPS base URL and the Bearer token, when required.

02

Set the identifier and test

Configure the endpoint, the query identifier and the name of the parameter expected by the API. Test with an existing record before proceeding.

03

Guide the trigger

In the activation step, describe when the connector should be triggered in the conversation. For example: fetch account data when the customer reports a technical issue and request the registered email if it’s not yet available.

04

Select the source in the agent

Enable and save the connector. In AI Agent, open Knowledge Base, select External connector via API, enable its use and save the changes.

05

Validate the support

Test conversations with and without the identifier, check the responses, and monitor calls in Recent activity. Adjust the guidelines to your API's purpose.

Your API controls the data. The agent handles the conversation.

A consulta acontece quando o contexto e as orientações configuradas pedem informações da fonte externa. Perguntas gerais e saudações continuam seguindo o atendimento do agente.

Access
Authorization in your systemThe provided identifier does not verify the customer's identity. Your API must enforce access rules and return only authorized data.
Data
Consultation during serviceThe API returns the information available in your system at that time. The update and availability of this data depend on the connected source.
Fonte
A source chosen by agentBy selecting API, this agent uses the external connector as its knowledge base. Files and Google Sheets will no longer be consulted by it.

What to know before connecting your API.

Do I need to set up an API?

Yes. Your system must provide a public HTTPS endpoint that receives the configured identifier and returns a JSON object or list with data. The team responsible for the system prepares the integration and its access rules.

Can I return different types of information?

Yes. The response can aggregate different authorized data into the same JSON structure. Each agent uses a connector; you can extend the data returned by your system's API.

Does the agent consult on every message?

No. It takes into account the conversation context and the guidance defined in When the connector should be triggered in the conversation. If it needs the identifier and it is not yet available, it requests that information from the customer.

Does the connector modify data in my system?

This feature performs GET requests to retrieve information. It is not a mechanism for creating, updating, or deleting records in the connected system.

What is the difference from Whatsplaid's public API?

Without an external connector, Whatsplaid queries your API to support conversations. In a public API setup, your system calls Whatsplaid to operate platform features. Explore the public API and webhooks.

Bring your system's data into the conversation.

Choose the knowledge source that fits your operation: external API, spreadsheets, or documents. Google Sheets · Document import

Start for free