Skip to main content
Weaver supports multiple messaging platforms through its channel integration system. Each channel enables bidirectional communication between users and AI agents across different platforms.

​
Supported Channels

Weaver currently supports the following messaging platforms:
  • Telegram - Popular messaging app with bot API support
  • Discord - Gaming and community chat platform
  • Slack - Professional workspace communication
  • LINE - Leading messaging platform in Asia
  • QQ - China’s most popular instant messaging platform

​
How Channels Work

Weaver’s channel architecture consists of:
  1. Channel Manager - Coordinates all active channels and message routing
  2. Message Bus - Central hub for inbound and outbound messages
  3. Channel Adapters - Platform-specific implementations

​
Message Flow

User Message → Platform API → Channel Adapter → Message Bus → Agent → Message Bus → Channel Adapter → Platform API → User

​
Configuration

Channels are configured in config.yaml under the channels section: 
channels:
  telegram:
    enabled: true
    token: "your-bot-token"
    allow_from:
      - "user_id_1"
      - "user_id_2"
  
  discord:
    enabled: true
    token: "your-bot-token"
    allow_from:
      - "user_id_1"

​
Common Configuration Options

All channels support these common settings:
FieldTypeDescription
enabledbooleanEnable/disable the channel
allow_fromarrayWhitelist of allowed user IDs (optional)

​
Access Control

Every channel implements an allowlist feature through the allow_from configuration:
  • If allow_from is empty or not set, all users can interact
  • If allow_from contains user IDs, only those users are allowed
  • User ID format varies by platform (see individual channel docs)

​
Media Support

Most channels support various media types:
  • Images - Automatically downloaded and processed
  • Voice/Audio - Can be transcribed using Groq Whisper
  • Files - Downloaded for agent processing
  • Video - Platform-dependent support

​
Voice Transcription

Channels can integrate with Groq’s Whisper API for voice message transcription: 
voice:
  enabled: true
  groq_api_key: "your-groq-key"
  model: "whisper-large-v3"
When enabled, voice messages are automatically transcribed and sent to the agent as text.

​
Starting Channels

The Channel Manager automatically starts all enabled channels when Weaver launches. Each channel:
  1. Initializes its connection to the platform API
  2. Registers event handlers for incoming messages
  3. Begins listening for user messages
  4. Subscribes to the outbound message bus

​
Message Metadata

Channels attach platform-specific metadata to each message:
  • message_id - Platform’s unique message identifier
  • user_id - Sender’s user ID
  • username - Sender’s display name
  • channel_id - Platform-specific channel/chat identifier
  • Additional platform-specific fields
This metadata helps agents provide context-aware responses.

​
Error Handling

All channels implement graceful error handling:
  • Failed message sends are logged but don’t crash the system
  • Network errors trigger automatic reconnection attempts
  • Invalid configurations prevent channel startup with clear error messages

​
Next Steps

Choose a platform to integrate:

Telegram

Set up Telegram bot integration

Discord

Configure Discord bot

Slack

Connect to Slack workspace

LINE

Integrate LINE Official Account

Build docs developers (and LLMs) love

Get started for freeTalk to us