Collections

Collections are the primary way to organize documents in Dexrag. Each collection acts as an isolated search space with its own adaptive learning model.

What is a Collection?

A collection is a group of related documents that share:

• Common search patterns
• Similar terminology
• Related use cases

Collection Modes

Standard Mode

Basic Monte Carlo RAG without adaptive learning. Great for static document sets.

const collection = await dex.create({
  name: 'my-docs',
  mode: 'standard'
})

Adaptive Mode (Recommended)

Learns from user behavior to improve over time. Personalized to your specific use case.

const collection = await dex.create({
  name: 'my-docs',
  mode: 'adaptive'
})

When to Create Multiple Collections

Create separate collections when you have:

1. Different Domains: Legal docs vs technical docs
2. Different Languages: English docs vs Spanish docs
3. Different Use Cases: Customer support vs internal knowledge base
4. Different Privacy Levels: Public vs confidential documents

Collection Limits

Best Practices

1. Use Descriptive Names

// Good
await dex.create({ name: 'legal-contracts-2024' })

// Bad
await dex.create({ name: 'coll1' })

2. Add Metadata

await dex.create({
  name: 'support-docs',
  description: 'Customer support documentation and FAQs',
  metadata: {
    department: 'support',
    language: 'en',
    version: '2.0'
  }
})

3. Monitor Performance

Track query performance and document relevance in the dashboard.

Deleting Collections

Deleting a collection removes:

• All documents
• Search history
• Adaptive learning data
• Analytics

This action cannot be undone.