1Create a Slack App

Navigate to https://api.slack.com/apps and create a new app. This is the app that you will later install into your Slack workspace to interact with the Dataiku Agent or LLM. You can configure the app yourself or use our pre-configured defaults using a manifest.

2Configuration Strategy

Use this manifest to create a pre-configured app. You can change the display names and descriptions to match your use case in the JSON, or configure that after the app creation.

View JSON manifest

{
    "display_information": {
        "name": "Dataiku Agent App",
        "description": "App to interact with a Dataiku Agent through a bot.",
        "background_color": "#1c2a38",
        "long_description": "I'm a bot to interact with a Dataiku Agent. You can talk to me in direct messages or mention me in a channel I'm part of. I will reply in the same thread or start one and consider the thread as part of our conversation."
    },
    "features": {
        "app_home": {
            "home_tab_enabled": true,
            "messages_tab_enabled": true,
            "messages_tab_read_only_enabled": false
        },
        "bot_user": {
            "display_name": "Dataiku Agent",
            "always_online": true
        },
        "assistant_view": {
            "assistant_description": "I'm a Dataiku Agent!",
            "suggested_prompts": []
        }
    },
    "oauth_config": {
        "scopes": {
            "bot": [
                "app_mentions:read",
                "channels:history",
                "channels:read",
                "chat:write",
                "groups:history",
                "groups:read",
                "im:history",
                "mpim:history",
                "reactions:write",
                "users:read",
                "users:read.email"
            ]
        }
    },
    "settings": {
        "event_subscriptions": {
            "bot_events": [
                "app_home_opened",
                "app_mention",
                "assistant_thread_started",                                
                "message.im"
            ]
        },
        "interactivity": {
            "is_enabled": true
        },
        "org_deploy_enabled": false,
        "socket_mode_enabled": true,
        "token_rotation_enabled": false
    }
}

To configure your app manually, these are the settings required for the integration to function correctly:

Enable Socket Mode:

In the left sidebar, select "Socket Mode"
Toggle "Enable Socket Mode" to ON if not enabled

Configure Event Subscriptions:

In the left sidebar, select "Event Subscriptions"
Under "Subscribe to bot events", select "Add Bot User Event" and add the following events:

Event Name	Description	Required Scope
app_home_opened	User opened the App Home on Slack	none
app_mention	Subscribe to only the message events that mention your app or bot	app_mentions:read
message.im	A message was posted in a direct message channel	im:history

These event subscriptions are required for your Slack bot to function properly.

3Generate an App Level Token

You will need to generate an app-level token to enable socket mode.

Go to api.slack.com/apps, find your Slack app for this integration if you're not already on the page
In the left sidebar select "Basic Information"
Navigate to App-Level Tokens and select "Generate Token and Scopes"
Enter a name for the token and add the connections:write scope
Copy the App-Level Token (starts with xapp-) to configure in the Dataiku Slack Webapp later

Thanks to Socket Mode, you do NOT need to configure the Events API URL nor expose your Dataiku instance.

Important

Each app-level token can only be used by one Dataiku WebApp on one Dataiku node. Sharing a token across multiple instances causes missing messages, as Slack randomly distributes events among connections.

4Install the Slack App to your Slack Workspace

After configuring the bot and permissions, select "Install App"
Select "Install to Your Workspace Name" to add your app to your Slack workspace
Authorize the requested permissions when prompted
Copy the Bot-User OAuth Token (starts with xoxb-) to configure in the Dataiku Slack Webapp later

Note

Since this installation uses socket mode, it is for internal workspace use only, and the Slack app cannot be publicly distributed.

5Create and Configure the Visual Webapp in Dataiku

In the webapp edit tab, configure the following:

Slack Authentication Settings:
- Bot User OAuth Token (From the Slack App Install App or OAuth & Permissions Menu)
- App-Level Token (From the Slack App Socket Mode Menu)
LLM: Select the Dataiku LLM to use for generating responses
Select "Save" to apply your configurations

Start the Backend:

When you save the webapp, the backend should automatically start
You'll see a notification indicating the backend is starting
If the backend doesn't start automatically:
- Go to the "Actions" panel on the right side of the screen
- Select "Start backend" to manually start it

6Test the Integration

In Slack:

Send a Direct Message to your bot
@mention your bot in a channel it has joined

In both cases, the bot should respond with a message from your configured Agent/LLM.

7Adding Bots to Channels

After installation, your bot does not automatically have access to all channels.
You must explicitly invite your bot to channels, e.g. using /invite @your_bot_name.
For the bot to be usable in private channels, your app needs the groups:history and groups:read permissions.
Unless added to a channel, the bot won't see messages or be able to respond in that channel.

Important

If you don't want your bot to be able to read data from private channels do not give it the groups:history and groups:read permissions.

These are the OAuth scopes required for your Slack app. Required scopes are necessary for core functionality, while optional scopes enable additional features.

Scope	Required	Purpose	Status
app_mentions:read	Required	View messages that directly mention your Slack app in conversations that the app is in	inactive
chat:write	Required	Enable the agent to send messages as your Slack app	inactive
users:read	Required	Enables the agent to read basic information about users, like their names. It's used to add context to queries to the agent.	inactive
channels:history	Recommended	View messages and other content in public channels that your Slack app has been added to. If not added, the agent will not be able to answer in public channels.	inactive
groups:history	Recommended	View messages and other content in private channels that your Slack app has been added to. If not added, the agent will not be able to answer in private channels.	inactive
im:history	Recommended	View messages and other content in direct messages that your Slack app has been added to. If not added, the agent will not be able to answer in direct private messages.	inactive
mpim:history	Recommended	View messages and other content in group direct messages that your Slack app has been added to. If not added, the agent will not be able to answer in group direct messages.	inactive
channels:read	Optional	Enables the agent to read the public channel name. This will be recorded in the LLM query, usable for later tracking.	inactive
groups:read	Optional	Enables the agent to read the private channel name. This will be recorded in the LLM query, usable for later tracking.	inactive
users:read.email	Optional	Enables the agent to get users email addresses. This is useful for LLM tracking, as the email will be recorded in LLM queries.	inactive
reactions:write	Optional	Allows the agent to use emojis (e.g., 🤔) to indicate task status.	inactive