Design AI search grounded in your documents.
What it does
Designs document retrieval systems with architecture diagrams, database trade-offs, retrieval pipelines, and evaluation plans.
When to use it
Use when building retrieval-augmented generation systems or knowledge-grounded applications that need semantic search and document retrieval.
How to use it
Give it a document retrieval system to design; it returns an architecture, database trade-offs, chunking rationale, retrieval pipeline, and evaluation plan.
Uses
Good to know
- cannot be undone: Qdrant
What you need · 2
The OpenAI embedding example supplies an API key to the client; the example does not establish where the actual key comes from.
The reranking implementation reads its Cohere API key from the COHERE_API_KEY environment variable.
Access · 3
This skill
Qdrant
Replaces a Qdrant collection (irreversible)
OpenAI
Sends document text to OpenAI