Skip to main content
The Whilst Slack bot accepts @mention queries, routes them to the appropriate data source, and responds with progressive updates and a final summarized answer.

​
How It Works

  1. Mention @whilst in any Slack channel with a question
  2. Bot sends an instant acknowledgement (within 3 seconds)
  3. Progressive updates show what’s happening (“Searching GitHub…”)
  4. Final answer posted as a threaded reply with citations
  5. Knowledge Pipeline runs in background to extract and document the conversation

​
Supported Data Sources

Whilst routes queries to the appropriate MCP adapter based on intent:
SourceStatusExample Query
GitHub✅ Live”What’s the latest open PR about churn?”
Linear✅ Live”What’s the status of the auth refactor?”
Notion🔜 Soon”Find the product spec for onboarding”
Whilst Docs✅ Live”What do we know about deployment?”

​
Progressive Updates

The bot uses Slack’s chat.update API to edit its original message as work progresses: 
🔍 Searching GitHub for relevant PRs...
    ↓
📊 Found 3 results, analyzing...
    ↓
✅ Here's what I found: [final answer with citations]
Rate limiting is handled via Slack API cadence controls with graceful degradation to a single final message on persistent 429s.

​
Multi-Tenant Architecture

Each Slack workspace has its own:
  • Bot token (encrypted at rest)
  • MCP credentials per data source
  • Job history and audit trail
  • Document corpus and embeddings
Credentials are resolved per-workspace from AWS Secrets Manager at /whilst/<env>/workspaces/<workspaceId>/....

​
Setup

See the Slack Setup Guide for installation and OAuth configuration.