# Question: Recommended integration approach for Node.js / TypeScript full-stack applications
Context
We are building a full-stack Next.js application with a Node.js / TypeScript backend. One of our core features is RAG-based document retrieval and question answering. We would like to use Ragas to evaluate retrieval quality and generated answers, including faithfulness, answer relevance, and context precision / recall.
Our application can provide the user question, retrieved document passages, generated answer, and reference answers where available. We are interested in both offline evaluation of curated datasets and asynchronous evaluation of sampled application interactions.
Questions
From the current installation instructions and quickstart, Ragas appears to be primarily distributed as a Python package. Could you clarify the recommended integration approach for an existing Node.js / TypeScript application?
- Is there an officially supported JavaScript / TypeScript SDK or another supported way to invoke the current Ragas evaluation APIs from Node.js?
- If Python is required, is the recommended architecture to run Ragas in a separate Python service or worker and exchange evaluation samples and results with the Next.js backend? Are there official examples or reference implementations for this setup?
- If native Node.js / TypeScript support is not currently available, is it planned or being tracked in an existing issue or discussion?
We would prefer to keep our application in TypeScript, but a separate Python evaluation component would also be an option if that is the supported approach.
Any guidance on the recommended integration boundary and relevant documentation would be appreciated. Thank you!
Source: vibrantlabsai/ragas