# How ngrok helps Sam develop and test LLM-powered RAG applications

DevFeed: [How ngrok helps Sam develop and test LLM-powered RAG applications](<https://devfeed.tech/articles/how-ngrok-helps-sam-develop-and-test-llm-powered-rag-applications-83320.md>)

Original publisher: [Read original article](<https://ngrok.com/blog/sam-develop-test-llm-rag>)

Author: Team ngrok

Published: 2025-02-06T00:00:00Z

Content type: article

Language: en

Sources: [\[ngrok news\]](<https://devfeed.tech/sources/ngrok-news.md>)

Topics: [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Holoscan for Media](<https://devfeed.tech/topics/holoscan-for-media.md>), [vercel ai sdk](<https://devfeed.tech/topics/vercel-ai-sdk.md>), [table driven tests](<https://devfeed.tech/topics/table-driven-tests.md>), [developer-productivity](<https://devfeed.tech/topics/developer-productivity.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [customers](<https://devfeed.tech/tags/customers.md>), [feature-flagging](<https://devfeed.tech/tags/feature-flagging.md>), [llm](<https://devfeed.tech/tags/llm.md>), [ngrok](<https://devfeed.tech/tags/ngrok.md>), [prompt](<https://devfeed.tech/tags/prompt.md>), [rag](<https://devfeed.tech/tags/rag.md>)

## AI overview

The guest post describes how Sam uses ngrok to develop distributed AI application services locally against staging systems, and to test prompt changes with a portion of production traffic without affecting production. Engineers compare the resulting local answers with production answers using a comparison service.

## Source excerpt

Building LLM-fueled apps? See how one AI startup uses ngrok to develop locally against remote services and safely test LLM prompt changes without breaking prod.