Skip to main content
The External Channels section enables you to connect your Agent with third-party providers so it can interact with users across popular messaging apps and service platforms. Unlike the Web and Bubble options, External Channels are tied to external providers (e.g., Meta, Freshchat, Zendesk), and each provider offers multiple channels (e.g., WhatsApp, Instagram, Messenger) with unique capabilities.

​
Integrate with a Provider

Available providers currently include: Once a provider is integrated, you can enable or disable channels individually.

​
Enable a Channel

Each provider supports one or more channels. After integration, you can enable or disable channels individually. Once enabled, the Agent will be able to send and receive messages through that channel.

​
Configure Add-ons

Each channel has add-ons that extend what the Agent can do. For example:
  • WhatsApp:
    • Send/receive voice notes
    • Enable voice recognition
    • Process image inputs (e.g., OCR, classification)
  • Messenger / Instagram:
    • Handle rich media messages (images, videos, stickers)
    • Add reaction handling
  • Freshchat:
    • Use callback URLs to link events back into your systems
Add-ons are optional and can be toggled on or off depending on the use case.
Important: Some add-ons may have additional costs depending on provider pricing or platform usage.

​
Configuration

When configuring a channel, you can:
  • Authenticate with the provider (e.g., connect to a Meta or Zendesk account).
  • Set permissions for the types of content your Agent can handle (voice, images, rich media).
  • Define behaviors that determine how the Agent should respond to different input types.

​
Conversation Behaviors (by Channel)

Each channel supports unique input and output methods. Examples include:
  • Text only (basic fallback for all channels).
  • Voice inputs/outputs (e.g., WhatsApp voice notes).
  • Image handling (user uploads → Agent processes).
  • File attachments (e.g., PDFs or documents shared in Zendesk-integrated chats).
Conversation behaviors are evolving — future releases will expand supported input/output formats and customization.

​
Example Scenarios

A user sends a voice note via WhatsApp.
  • The Agent can process it via voice recognition.
  • The Agent may reply with either text or another voice note.
  • Optional add-ons: image recognition, OCR for attachments.
A user shares an image via Instagram.
  • The Agent can classify or extract information from the image.
  • Replies may include text, rich media, or contextual information.
  • Optional add-ons: advanced media processing, stickers, reactions.
A user reacts to a message with an emoji in Messenger.
  • The Agent acknowledges the reaction.
  • The Agent may adapt its next response based on the sentiment or context.
  • Optional add-ons: handling rich media (images, videos).