Skip to main content

Documentation Index

Fetch the complete documentation index at: https://simili.mintlify.app/llms.txt

Use this file to discover all available pages before exploring further.

​
Qdrant Integration

How Simili Bot uses Qdrant for semantic search and vector storage.

​
Overview

Qdrant is the vector database backend for Simili Bot. It stores:
  • Issue embeddings (3072-dimensional vectors)
  • Issue metadata (title, repo, number, labels, etc.)
  • Enables fast semantic similarity search

​
Key features

Collection Management
  • Automatic collection creation
  • Configurable collection names
  • Payload storage for metadata
Search Capabilities
  • Cosine similarity scoring
  • Configurable thresholds
  • Fast retrieval (100ms-1s typical)
Scalability
  • Handles thousands to millions of vectors
  • Efficient indexing and retrieval
  • Cloud or self-hosted options

​
Configuration

qdrant:
  url: "https://cluster.qdrant.io:6333"
  api_key: "${QDRANT_API_KEY}"
  collection: "issues"
  tls: true
  timeout: 30
  max_retries: 3

​
Storage

Per-issue vector storage: 
Vector: 768 dimensions (float32)
Metadata: ~1-2KB (title, body, labels, etc.)
Total: ~3-4KB per issue
For 10,000 issues: ~30-40MB of storage

​
Performance

Typical Operations:
  • Point insertion: 100-500ms
  • Similarity search: 500ms-1s
  • Batch operations: Parallelizable
Scaling:
  • Linear with vector count
  • Optimized for cosine distance
  • Supports fuzzy search

​
Deployment

Qdrant Cloud (Recommended)
  • Managed service
  • Automatic backups
  • 1GB free tier
  • Pay-as-you-go pricing
Self-Hosted
  • Docker container
  • Docker Compose
  • Kubernetes deployment
  • Full control

​
Integration points

  1. Indexing - Add issues to Qdrant
  2. Search - Find similar issues
  3. Updates - Re-index modified issues
  4. Cleanup - Archive old issues

​
Next steps

Qdrant configuration

Configure Qdrant connection

Semantic search

How semantic search works