Search documentation

Search for docs or ask AI

Data Sources Overview

Pillar's co-pilot uses your knowledge base to answer user questions. This section covers how to connect external data sources to keep your content in sync.

How It Works

Your Content              Pillar                     User
┌─────────────┐      ┌──────────────┐      ┌─────────────────┐
│ GCS Bucket  │      │              │      │                 │
│ S3 Bucket   │─────▶│  Knowledge   │─────▶│    Co-pilot     │
│ Website     │      │  Base Sync   │      │  (Chat Panel)   │
│ Help Center │      │              │      │                 │
└─────────────┘      └──────────────┘      └─────────────────┘
  1. You configure a data source in the Pillar dashboard
  2. Pillar syncs your content on a schedule or on-demand
  3. Content is processed and indexed for AI retrieval
  4. Users ask questions and get answers from your knowledge

Supported Sources

Source TypeDescriptionSetup Guide
Website CrawlAutomatically crawl and index your website or docsDashboard
Google Cloud StorageSync documents from a GCS bucketGCS Setup
Amazon S3Sync documents from an S3 bucketS3 Setup
Direct UploadUpload files directly in the dashboardDashboard

Sync Behavior

Automatic Sync

By default, Pillar syncs your content:

  • Immediately when you first connect a source
  • Daily to pick up new or updated content
  • On-demand when you trigger a manual sync

Incremental Updates

Pillar uses smart diffing to:

  • Only process new or changed documents
  • Remove content that's been deleted from the source
  • Preserve any manual corrections you've made

Supported File Types

TypeExtensions
Documents.md, .mdx, .txt, .pdf, .docx
Web Pages.html, .htm
Structured.json, .yaml, .yml

Best Practices

Organize Your Content

  • Use clear, descriptive file names
  • Group related content in folders
  • Include metadata (title, description) in your documents

Keep Content Fresh

  • Set up automatic syncs for frequently updated content
  • Use webhooks to trigger syncs after deployments
  • Review sync logs to catch any issues

Optimize for AI

  • Write clear, concise content
  • Use headings to structure information
  • Include examples and code snippets where relevant

Next Steps