AgentShield for LlamaIndex

Instrument LlamaIndex RAG pipelines and agents. Track retrieval costs, monitor query latency, protect with PII redaction.

Quick setup

import agentshield
from llama_index.core import VectorStoreIndex

shield = agentshield.init(api_key="your-key")

with shield.session(agent_id="rag-pipeline") as session:
    index = VectorStoreIndex.from_documents(docs)
    query_engine = index.as_query_engine()
    response = query_engine.query(user_query)
    session.track(response=str(response))