Skip to main content

Use Slack with SmythOS

Connect your agents to Slack for real-time team communication, automated alerts, and contextual interactions via components or Model Context Protocol (MCP).

TL;DR
Authenticate your Slack app, configure the Send Message component, or deploy using Model Context Protocol (MCP) to route intelligent agent outputs into Slack.

Why Use Slack With Your Agent?​

Slack is a core communication tool for modern teams. Integrating it with SmythOS allows you to:

  • Trigger alerts, summaries, or updates from automated workflows
  • Send messages based on logic, triggers, or API events
  • Send alerts or summaries directly to teams
  • Build chatOps flows for faster task execution
  • Route responses from AI agents into Slack threads
  • Customize interactions based on channel, user, or context

With both direct Slack component integration and Model Context Protocol (MCP), you can choose how and where Slack fits in.

Prerequisites​

Make sure you have:

  • A Slack Workspace and Slack App
  • Bot User enabled with messaging scopes
  • Your Bot User OAuth Token
  • Optional: Store the token in Vault → API Keys

Option 1: Send Slack Messages Using the Component​

Step 1: Get Your Slack Bot Token​

  1. Create a Slack App at Slack API Console
  2. Enable the Bot User under App Features
  3. Install the app under OAuth & Permissions
  4. Copy your Bot User OAuth Token
  5. Add it manually or store it in Vault
INFO
Use Vault for secure, multi-agent reuse and better token management.

Step 2: Configure the Send Message Component​

Open the Slack Send Message component in your workflow.

  • Set authentication to use Vault key or paste token
  • Define inputs for your message
{
  "channel_id": "C12345678",
  "text": "Your agent has completed the task."
}

Inputs Required:

  • channel_id: ID of the Slack channel
  • text: Message content (≤ 2000 characters)

Step 3: Output Structure​

Once the message sends, the outputs include:

  • response: Full Slack message object
  • headers: Metadata from Slack API
  • status: HTTP status code
  • msgObj: Message ID, content, timestamp

Option 2: Deploy Slack Integration via Model Context Protocol (MCP)​

If your agent uses memory, roles, or multi-modal workflows, deploy with MCP to:

  • Route outputs to Slack after deeper context processing
  • Combine Slack notifications with email, Discord, or logging
  • Enable multi-agent collaboration across departments

Sample Payload to MCP:

{
  "input": "New ticket submitted in support.",
  "platform": "slack",
  "channel_id": "C024BE91L"
}

Follow this with a Send Message component for delivery.

INFO
MCP is best for routing Slack messages after memory/context-based decision logic.

What’s Next?​

Connect your SmythOS agent to Slack using a secure bot token and deliver automated updates with the Send Message component or via MCP.