RAG - PraisonAI

RAG (Retrieval-Augmented Generation) enhances agents with external knowledge.

RAG::query is not yet implemented in the Rust SDK — calling it returns an Error::Config ("RAG::query is not implemented..."). The surrounding types (chunking, context packing, citations, retrieval configuration) are real and usable, and knowledge-backed agents via Agent::knowledge(KnowledgeConfig) still work. Track full retrieval support in PraisonAI#4944.

Quick Start

Set Up RAG

use praisonai::{Agent, KnowledgeConfig};

let config = KnowledgeConfig::new()
    .source("docs/")
    .chunk_size(1000)
    .retrieval_k(5);

let agent = Agent::new()
    .name("RAG Bot")
    .knowledge(config)
    .build()?;

agent.chat("Answer based on documents").await?;

RAG Components

Component	Description	Status
Loader	Load documents	✅ documented
Chunker	Split into chunks	✅ documented
Embedder	Create vectors	✅ documented
Retriever	Find relevant chunks	✅ documented
Reranker	Improve relevance	✅ documented
Query pipeline (`RAG::query`)	End-to-end retrieval + generation	⚠️ not implemented — returns `Err`

Knowledge

Knowledge base

Chunking

Document chunking

​Quick Start

​RAG Components

​Related

Knowledge

Chunking

Quick Start

RAG Components

Related