Channels
Tavio supports 10 communication channels. Every channel feeds into the same unified AI pipeline, so your knowledge base, agent configurations, and analytics work identically across all of them.
Unified Pipeline
Regardless of the channel, every inbound message follows the same path:
Channel Adapter -> Message Router -> Conversation Service -> RAG Retriever -> AI Engine -> Response
Each channel adapter normalizes the incoming format and enforces channel-specific constraints (character limits, media support, session handling). The AI engine receives a clean, uniform message regardless of origin.
Token Limits by Channel
AI responses are capped per channel to fit delivery constraints:
| Channel | Max Tokens | Reason |
|---|---|---|
| Widget | 1000 | No delivery constraint |
| 500 | Message length best practice | |
| Voice | 150 | Spoken response brevity |
| SMS | 140 | SMS segment limits |
| USSD | 130 | Screen size constraint |
All Channels
Embeddable chat widget for websites and web apps.
Max tokens: 1000
Meta Cloud API integration for WhatsApp Business.
Max tokens: 500
Africa's Talking and Twilio SMS delivery.
Max tokens: 140
Interactive menu-based sessions via Africa's Talking.
Max tokens: 130
Twilio Voice with ElevenLabs TTS for IVR flows.
Max tokens: 150
Telegram Bot API with webhook-based messaging.
Max tokens: 500
Meta Send/Receive API for Messenger conversations.
Max tokens: 500
Instagram DM automation via Meta Graph API.
Max tokens: 500
X API v2 for auto-reply drafting on mentions.
Max tokens: 500
Discord bot integration for community support.
Max tokens: 500