# \#rag

**URL:** https://community.openai.com/tag/rag/268.md

[Latest](https://community.openai.com/latest.md) · [Categories](https://community.openai.com/categories.md) · [Tags](https://community.openai.com/tags.md)

---

## [What We've Learned Building RAG Pipelines That Don't Fall Apart at Scale](https://community.openai.com/t/what-weve-learned-building-rag-pipelines-that-dont-fall-apart-at-scale/1395780)

<div class="topic-metadata">

**Author:** [@divisha.jain](https://community.openai.com/u/divisha.jain)\
**Replies:** 0\
**Last updated:** [September 8, 2026, 11:01am UTC](https://community.openai.com/t/what-weve-learned-building-rag-pipelines-that-dont-fall-apart-at-scale/1395780 "2026-09-08T11:01:30Z")

</div>

RAG looks simple in a tutorial: chunk your docs, embed them, retrieve top-k, stuff into context, done. That setup works fine for a demo with 50 documents. It falls apart fast once you’re dealing with tens of thousands of…

---

## [VERA – Portable Document RAG for Humans and Agents](https://community.openai.com/t/vera-portable-document-rag-for-humans-and-agents/1392642)

<div class="topic-metadata">

**Author:** [@willis8](https://community.openai.com/u/willis8)\
**Replies:** 0\
**Last updated:** [August 26, 2026, 3:51am UTC](https://community.openai.com/t/vera-portable-document-rag-for-humans-and-agents/1392642 "2026-08-26T03:51:43Z")

</div>

Retrieval-Augmented Generation (RAG) has become one of the most useful machine learning technologies in my daily work as a civil engineer, even as LLM context windows continue to grow. That’s why I created VERA (Vector E…

---

## [Which embedding model works best for RAG when using GPT-5.4 mini/nano or GPT-4.1 mini/nano as the LLM?](https://community.openai.com/t/which-embedding-model-works-best-for-rag-when-using-gpt-5-4-mini-nano-or-gpt-4-1-mini-nano-as-the-llm/1386850)

<div class="topic-metadata">

**Author:** [@meharaz733](https://community.openai.com/u/meharaz733)\
**Replies:** 5\
**Last updated:** [July 16, 2026, 5:10pm UTC](https://community.openai.com/t/which-embedding-model-works-best-for-rag-when-using-gpt-5-4-mini-nano-or-gpt-4-1-mini-nano-as-the-llm/1386850 "2026-07-16T17:10:06Z")

</div>

Hi everyone, I’m building a Retrieval-Augmented Generation (RAG) system and I’m trying to choose the right embedding model. My generation models will likely be one of these: GPT-5.4 mini / nano GPT-4.1 mini / nan…

---

## [My AI Questions and Hopefully answers](https://community.openai.com/t/my-ai-questions-and-hopefully-answers/1383161)

<div class="topic-metadata">

**Author:** [@yash7298w](https://community.openai.com/u/yash7298w)\
**Replies:** 8\
**Last updated:** [June 22, 2026, 10:06am UTC](https://community.openai.com/t/my-ai-questions-and-hopefully-answers/1383161 "2026-06-22T10:06:12Z")

</div>

Retrieval-Augmented Generation (RAG) is widely used in LLM-based applications to connect models with external knowledge sources. It’s often presented as the default solution for improving factual accuracy and grounding r…

---

## [How do you forecast what an AI feature will actually cost — before you build it?](https://community.openai.com/t/how-do-you-forecast-what-an-ai-feature-will-actually-cost-before-you-build-it/1382064)

<div class="topic-metadata">

**Author:** [@shokhrukhkarimov](https://community.openai.com/u/shokhrukhkarimov)\
**Replies:** 12\
**Last updated:** [October 30, 2026, 5:13pm UTC](https://community.openai.com/t/how-do-you-forecast-what-an-ai-feature-will-actually-cost-before-you-build-it/1382064 "2026-10-30T17:13:22Z")

</div>

When I shipped my first AI feature, I budgeted by gut feel, and the real bill at scale was nothing like my estimate — re-embedding, retries, and a too-expensive default model quietly added up. I’d have made very differen…

---

## [RAG is not really a solution](https://community.openai.com/t/rag-is-not-really-a-solution/599291)

<div class="topic-metadata">

**Author:** [@joyasree78](https://community.openai.com/u/joyasree78)\
**Replies:** 113\
**Last updated:** [April 25, 2026, 7:17pm UTC](https://community.openai.com/t/rag-is-not-really-a-solution/599291 "2026-04-25T19:17:16Z")

</div>

After spending more than a year now with Gen AI, I feel RAG is more of a problem than a solution. it is so brittle and there is no science to it. I wanted to check with this group if anyone is aware of any other techniqu…

---

## [Vector Storage RAG - AI Agent Not Retrieving Data Beyond Midpoint of Uploaded CSV](https://community.openai.com/t/vector-storage-rag-ai-agent-not-retrieving-data-beyond-midpoint-of-uploaded-csv/1378766)

<div class="topic-metadata">

**Author:** [@nbo2](https://community.openai.com/u/nbo2)\
**Replies:** 7\
**Last updated:** [April 12, 2026, 4:09am UTC](https://community.openai.com/t/vector-storage-rag-ai-agent-not-retrieving-data-beyond-midpoint-of-uploaded-csv/1378766 "2026-04-12T04:09:10Z")

</div>

Hello everyone, I’m encountering an issue reported by a customer and was able to reproduce it myself. The scenario is as follows: A CSV file (~1MB) with ~500 rows and 2 simple columns (SKU, Inventory) is uploaded. …

---

## [From GEO to DED: Governing How AI Represents Brands](https://community.openai.com/t/from-geo-to-ded-governing-how-ai-represents-brands/1373163)

<div class="topic-metadata">

**Author:** [@sedaefe](https://community.openai.com/u/sedaefe)\
**Replies:** 0\
**Last updated:** [February 2, 2026, 7:49pm UTC](https://community.openai.com/t/from-geo-to-ded-governing-how-ai-represents-brands/1373163 "2026-02-02T19:49:56Z")

</div>

I’ve been working on a problem I keep seeing across LLM-based systems: models can retrieve the “right” information, but still represent it in the wrong context. I define this gap as: GEO (Generative Engine Optimizatio…

---

## [Strategies to summarize succinctly?](https://community.openai.com/t/strategies-to-summarize-succinctly/1371998)

<div class="topic-metadata">

**Author:** [@whafa](https://community.openai.com/u/whafa)\
**Replies:** 7\
**Last updated:** [January 19, 2026, 6:04pm UTC](https://community.openai.com/t/strategies-to-summarize-succinctly/1371998 "2026-01-19T18:04:10Z")

</div>

Hi, if this isn’t the right place to talk about this sort of thing, could you help me find that place? Thanks. I’m working with (IMO) a large set of message board data. I’ve got my ETL pipeline down to the point where…

---

## [OpenAI Realtime API w/ Twilio + RAG == AI Call Center](https://community.openai.com/t/openai-realtime-api-w-twilio-rag-ai-call-center/981632)

<div class="topic-metadata">

**Author:** [@alden](https://community.openai.com/u/alden)\
**Replies:** 8\
**Last updated:** [December 22, 2025, 11:55pm UTC](https://community.openai.com/t/openai-realtime-api-w-twilio-rag-ai-call-center/981632 "2025-12-22T23:55:56Z")

</div>

I’d like to share this POC that connects the OpenAI Realtime API to a Twilio phone number – and then – augments it with RAG knowledge. So something like this lets you “Talk to Website” - yeah, literally calling a phone …

---

## [Research Survey on RAG Development Practices & Challenges (8-10 mins)](https://community.openai.com/t/research-survey-on-rag-development-practices-challenges-8-10-mins/1366570)

<div class="topic-metadata">

**Author:** [@KHChan](https://community.openai.com/u/KHChan)\
**Replies:** 2\
**Last updated:** [November 17, 2025, 7:45am UTC](https://community.openai.com/t/research-survey-on-rag-development-practices-challenges-8-10-mins/1366570 "2025-11-17T07:45:01Z")

</div>

Hey everyone! :waving\_hand: I’m a final-year CS student at NTU Singapore working on my Final Year Research about RAG pipelines. I’m working on a tool designed to help developers evaluate and compare different RAG techni…

---

## [How to Reduce LLM Confidence So It Always Queries the VectorStore?](https://community.openai.com/t/how-to-reduce-llm-confidence-so-it-always-queries-the-vectorstore/1366318)

<div class="topic-metadata">

**Author:** [@ibiscp](https://community.openai.com/u/ibiscp)\
**Replies:** 1\
**Last updated:** [November 14, 2025, 7:14pm UTC](https://community.openai.com/t/how-to-reduce-llm-confidence-so-it-always-queries-the-vectorstore/1366318 "2025-11-14T19:14:18Z")

</div>

I’m building a project that uses a VectorStore, and I’m running into a problem with model confidence. The issue is that the LLM is often too confident in its own internal knowledge. I want it to be less confident so tha…

---

## [Openai vector stores -- is chunking still required or does openai do it for me?](https://community.openai.com/t/openai-vector-stores-is-chunking-still-required-or-does-openai-do-it-for-me/1366213)

<div class="topic-metadata">

**Author:** [@Tom\_Gagne](https://community.openai.com/u/Tom_Gagne)\
**Replies:** 1\
**Last updated:** [November 13, 2025, 11:12pm UTC](https://community.openai.com/t/openai-vector-stores-is-chunking-still-required-or-does-openai-do-it-for-me/1366213 "2025-11-13T23:12:31Z")

</div>

I’ve read competing documentation on whether chunking is required. I’ve read that it used to be required but that documents are automatically chunked now when loaded into vector-stores. Is that true only when loaded vi…

---

## [How to extend white list for Files API (add .swift, .mm, .m)](https://community.openai.com/t/how-to-extend-white-list-for-files-api-add-swift-mm-m/1362713)

<div class="topic-metadata">

**Author:** [@macsis](https://community.openai.com/u/macsis)\
**Replies:** 0\
**Last updated:** [October 16, 2025, 9:46am UTC](https://community.openai.com/t/how-to-extend-white-list-for-files-api-add-swift-mm-m/1362713 "2025-10-16T09:46:32Z")

</div>

We use Vector Stores (and Files API) for retrieval over our codebase. Our repositories contain many Swift and ObjC files. Currently, the Files API rejects .swift/.m/.mm files with “File type not supported” (HTTP 400), wh…

---

## [Avoid explicit mention to retrieval and assistant files](https://community.openai.com/t/avoid-explicit-mention-to-retrieval-and-assistant-files/565091)

<div class="topic-metadata">

**Author:** [@mmisol](https://community.openai.com/u/mmisol)\
**Replies:** 12\
**Last updated:** [October 13, 2025, 3:49pm UTC](https://community.openai.com/t/avoid-explicit-mention-to-retrieval-and-assistant-files/565091 "2025-10-13T15:49:59Z")

</div>

Hi everyone! I have an assistant with the retrieval tool and some files associated with it. I want the assistant to use the information in the files for RAG but this should be done behind the scenes. What I see is that …

---

## [Mitigating Hallucinations in RAG - a 2025 review](https://community.openai.com/t/mitigating-hallucinations-in-rag-a-2025-review/1362063)

<div class="topic-metadata">

**Author:** [@GenerativeJourney](https://community.openai.com/u/GenerativeJourney)\
**Replies:** 7\
**Last updated:** [October 12, 2025, 2:40am UTC](https://community.openai.com/t/mitigating-hallucinations-in-rag-a-2025-review/1362063 "2025-10-12T02:40:46Z")

</div>

Stumbled upon this review while doing literature research on LLM hallucinations. Recommend reading if you are working with RAG. https://www.mdpi.com/2227-7390/13/5/856

---

## [Functions vs. Classification: Best Practice for selecting an action based on User Intent](https://community.openai.com/t/functions-vs-classification-best-practice-for-selecting-an-action-based-on-user-intent/1355964)

<div class="topic-metadata">

**Author:** [@sol-ruh-1](https://community.openai.com/u/sol-ruh-1)\
**Replies:** 0\
**Last updated:** [September 3, 2025, 7:32am UTC](https://community.openai.com/t/functions-vs-classification-best-practice-for-selecting-an-action-based-on-user-intent/1355964 "2025-09-03T07:32:53Z")

</div>

Hello everyone, I’m building a RAG system where a user asks a question, and I need to determine whether they’re asking for a single, specific file or a list of related files from a database. I’m exploring two main archi…

---

## [Understanding Grounding Under the Hood](https://community.openai.com/t/understanding-grounding-under-the-hood/1345174)

<div class="topic-metadata">

**Author:** [@miro7](https://community.openai.com/u/miro7)\
**Replies:** 7\
**Last updated:** [August 22, 2025, 1:14am UTC](https://community.openai.com/t/understanding-grounding-under-the-hood/1345174 "2025-08-22T01:14:40Z")

</div>

I am trying to get a clearer picture of how grounding works under the hood in large language models. I have two questions and I hope someone can help me to clarify. 1- From what I understand, RAG is the only mechanism u…

---

## [How to optimize chunked grammar/spell-check processing with LLaMA.cpp in Node.js?](https://community.openai.com/t/how-to-optimize-chunked-grammar-spell-check-processing-with-llama-cpp-in-node-js/1352978)

<div class="topic-metadata">

**Author:** [@Sanket\_Jaiswal](https://community.openai.com/u/Sanket_Jaiswal)\
**Replies:** 1\
**Last updated:** [August 21, 2025, 4:08pm UTC](https://community.openai.com/t/how-to-optimize-chunked-grammar-spell-check-processing-with-llama-cpp-in-node-js/1352978 "2025-08-21T16:08:02Z")

</div>

Hi everyone, I’m working on a custom RAG-like system in Node.js where users can choose options like grammar correction or spell checking. To avoid hitting the model’s token limit, I split large documents into smaller ch…

---

## [I built a personal news agent using openai embedding model](https://community.openai.com/t/i-built-a-personal-news-agent-using-openai-embedding-model/1348970)

<div class="topic-metadata">

**Author:** [@Vince\_Yip](https://community.openai.com/u/Vince_Yip)\
**Replies:** 1\
**Last updated:** [August 20, 2025, 8:06pm UTC](https://community.openai.com/t/i-built-a-personal-news-agent-using-openai-embedding-model/1348970 "2025-08-20T20:06:38Z")

</div>

I built a news agent that helps you easily follow any topic. You just type in what you want to follow, AI keeps fetching the latest news for you every hour. How it works: Ingests 2,000+ RSS feeds hourly (The Verge, NY…

---

## [Filtering by metadata for File Search workaround](https://community.openai.com/t/filtering-by-metadata-for-file-search-workaround/1344113)

<div class="topic-metadata">

**Author:** [@marcinc](https://community.openai.com/u/marcinc)\
**Replies:** 3\
**Last updated:** [August 14, 2025, 7:34am UTC](https://community.openai.com/t/filtering-by-metadata-for-file-search-workaround/1344113 "2025-08-14T07:34:13Z")

</div>

We make use of assistant, vector store and file search. We process and ingest contracts data for different companies and questions asked by end users are always within the context of the company and we want to filter re…

---

## [Responses API file\_search tool - issues and guidance](https://community.openai.com/t/responses-api-file-search-tool-issues-and-guidance/1210922)

<div class="topic-metadata">

**Author:** [@elliotfraser](https://community.openai.com/u/elliotfraser)\
**Replies:** 8\
**Last updated:** [July 2, 2025, 10:33am UTC](https://community.openai.com/t/responses-api-file-search-tool-issues-and-guidance/1210922 "2025-07-02T10:33:44Z")

</div>

Hey everyone, I’m working with the new responses API and trying out the out-of-the-box file\_search tool. Here’s what I’ve done so far: Set up a vector store and added the file\_search tool schema File search is working,…

---

## [How to build an AI system that can search over 50,000 documents with high accuracy?](https://community.openai.com/t/how-to-build-an-ai-system-that-can-search-over-50-000-documents-with-high-accuracy/1285970)

<div class="topic-metadata">

**Author:** [@francotrigo](https://community.openai.com/u/francotrigo)\
**Replies:** 6\
**Last updated:** [June 13, 2025, 7:52pm UTC](https://community.openai.com/t/how-to-build-an-ai-system-that-can-search-over-50-000-documents-with-high-accuracy/1285970 "2025-06-13T19:52:14Z")

</div>

Hi everyone, I’m working on a project that involves building an AI system capable of answering natural language questions based on a large collection of documents—more than 50,000 legal and regulatory files (laws, contr…

---

## [Memory Full in GPT-4 Assistant – How to Safely Clean Up Without Losing AI Knowledge?](https://community.openai.com/t/memory-full-in-gpt-4-assistant-how-to-safely-clean-up-without-losing-ai-knowledge/1282006)

<div class="topic-metadata">

**Author:** [@Szilagyi\_Peter](https://community.openai.com/u/Szilagyi_Peter)\
**Replies:** 2\
**Last updated:** [June 9, 2025, 6:01pm UTC](https://community.openai.com/t/memory-full-in-gpt-4-assistant-how-to-safely-clean-up-without-losing-ai-knowledge/1282006 "2025-06-09T18:01:01Z")

</div>

Title: Memory Full in GPT-4 Assistant – How to Safely Clean Up Without Losing AI Knowledge? Hi everyone, I’m developing a business-critical custom AI assistant using ChatGPT Plus with long-term memory enabled. The assi…

---

## [How about adding a new role for context in RAG applications](https://community.openai.com/t/how-about-adding-a-new-role-for-context-in-rag-applications/1260735)

<div class="topic-metadata">

**Author:** [@jordimarti](https://community.openai.com/u/jordimarti)\
**Replies:** 5\
**Last updated:** [June 3, 2025, 5:19am UTC](https://community.openai.com/t/how-about-adding-a-new-role-for-context-in-rag-applications/1260735 "2025-06-03T05:19:52Z")

</div>

In Retrieval-Augmented Generation (RAG) applications, it’s not always clear where or how to include context information within the current OpenAI API structure. I think it would be helpful to introduce a new role — somet…

---

## [Best approaches for querying JSON files with LLMs](https://community.openai.com/t/best-approaches-for-querying-json-files-with-llms/1273008)

<div class="topic-metadata">

**Author:** [@Enzo26](https://community.openai.com/u/Enzo26)\
**Replies:** 5\
**Last updated:** [May 30, 2025, 12:07pm UTC](https://community.openai.com/t/best-approaches-for-querying-json-files-with-llms/1273008 "2025-05-30T12:07:03Z")

</div>

I’m trying to access a json file with 56 objects each with key/values that i want my assistant or response api to query for specific user selection inputs as well as short text inputs that I’d like for the model to reas…

---

## [Implementing RAG via Custom Functions in OpenAI Assistants](https://community.openai.com/t/implementing-rag-via-custom-functions-in-openai-assistants/984516)

<div class="topic-metadata">

**Author:** [@abc01](https://community.openai.com/u/abc01)\
**Replies:** 10\
**Last updated:** [May 10, 2025, 9:43am UTC](https://community.openai.com/t/implementing-rag-via-custom-functions-in-openai-assistants/984516 "2025-05-10T09:43:06Z")

</div>

Hi everyone, I’m exploring the possibility of implementing Retrieval-Augmented Generation (RAG) directly using the “custom functions” feature in OpenAI Assistants. Specifically, given a vector storage database (such as…

---

## [My metadata isn't correctly processed when uploading a JSON file](https://community.openai.com/t/my-metadata-isnt-correctly-processed-when-uploading-a-json-file/1253073)

<div class="topic-metadata">

**Author:** [@LUNY](https://community.openai.com/u/LUNY)\
**Replies:** 2\
**Last updated:** [May 6, 2025, 12:01pm UTC](https://community.openai.com/t/my-metadata-isnt-correctly-processed-when-uploading-a-json-file/1253073 "2025-05-06T12:01:37Z")

</div>

Hello everyone, I’m working on a RAG system, I want to store my chunks with some additional metadata, to do so I used a json file. Here is a snippet of my code: for pdf in tqdm(list(Path(".").glob("\*.pdf")), unit="pdf"…

---

## [Clarification on Specifying Embedding Models in OpenAI Vector Stores](https://community.openai.com/t/clarification-on-specifying-embedding-models-in-openai-vector-stores/1241740)

<div class="topic-metadata">

**Author:** [@gopichandar](https://community.openai.com/u/gopichandar)\
**Replies:** 0\
**Last updated:** [April 25, 2025, 8:39am UTC](https://community.openai.com/t/clarification-on-specifying-embedding-models-in-openai-vector-stores/1241740 "2025-04-25T08:39:21Z")

</div>

I’m seeking clarification regarding the specification of embedding models when creating vector stores using the OpenAI API. According to the official documentation for the POST /v1/vector-stores endpoint, there is no me…

---

## [Seeking help from the experts about improvement of GraphRAG Drift Search!](https://community.openai.com/t/seeking-help-from-the-experts-about-improvement-of-graphrag-drift-search/1230143)

<div class="topic-metadata">

**Author:** [@1372417922](https://community.openai.com/u/1372417922)\
**Replies:** 1\
**Last updated:** [April 16, 2025, 8:08am UTC](https://community.openai.com/t/seeking-help-from-the-experts-about-improvement-of-graphrag-drift-search/1230143 "2025-04-16T08:08:28Z")

</div>

While studying the Drift Search mechanism in GraphRAG, I observed a potential efficiency issue related to entity redundancy. Here’s my analysis: Redundancy in Sub-queries (in drift search): When configuring the to…

[Next page](https://community.openai.com/tag/rag/268.md?match_all_tags=true&page=1&tags%5B%5D=rag)
