Workflow Operator (Chat Bot)
Workflow Operator brings the built-in Dagu agent into Slack, Telegram, Discord, and LINE. Each channel, thread, DM, or LINE source keeps its own ongoing conversation, so you can ask follow-up questions and receive run notifications in the same place.
Only one messaging connector can be active at a time. Choose it with bots.provider in the config file or DAGU_BOTS_PROVIDER in the environment.
Prerequisite
Configure AI agent first. Go to AI agent Settings (/agent-settings) in the Web UI and set your model, tool policy, and related defaults. Start with AI agent Settings, then use Models & Providers and Tool Permissions & Bash Policy as needed.
Available Platforms
| Platform | Connection | Description |
|---|---|---|
| Telegram | Long polling | Use Workflow Operator from Telegram chat, receive notifications, and continue follow-up in the same conversation |
| Slack | Socket Mode (WebSocket) | Use Workflow Operator from Slack channels and threads for operational follow-up without leaving chat |
| Discord | Gateway (WebSocket) | Use Workflow Operator from Discord channels and DMs with button-driven prompts and inline notifications |
| LINE | Messaging API webhook | Use Workflow Operator from LINE one-on-one chats, groups, and rooms. Requires a public HTTPS webhook endpoint or a managed instance. |
Relationship to AI agent
Workflow Operator is built on top of the same built-in steward used in the Web UI.
That means:
- AI agent must be enabled first
- A working default model must exist
- Tool permissions and bash policy still apply
- Provider-specific model issues such as local-model endpoint configuration still matter
Start with AI agent Settings, then return here to configure Slack, Telegram, Discord, or LINE.