AutoChunks

Embedding & Retrieval Shared Backbone

Documents Path

Scan Results

Enable Proxy Sampling

Optimization Mode

Intelligence

Embedding Provider

Model ID or Local Path

Optimization Goal

Enable Tracing (Arize Phoenix)

Enable RAGAS Analysis

Retrieval via: Hashing (Mock)

Strategy & Search

Native Optimizers

Fixed Length Recursive Char Semantic Local Layout Aware Hybrid Parent-Child Sentence Aware HTML Section

LangChain Bridges

LC Recursive LC Character LC Markdown LC Token LC Python LC HTML LC JSON

Chunk Sizes (Tokens)

256 512 1024

Overlap Ratio

10% 15% 20%

Control Center