Additionally, there might be a potential route where both are used, depending on the use case.
Feel free to dm if you want to chat further on this!"},"story_title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["rag","framework"],"value":"Show HN: Neum AI \u2013 Open-source large-scale RAG framework"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/NeumTry/NeumAI"}},"_tags":["comment","author_picohen","story_38368570"],"author":"picohen","children":[38372340,38377088],"comment_text":"Well, of course I'm biased on the answer :). But to give a not-so-biased answer, I would first try to understand what the project is about and whether RAG is a priority in it.\nIf the project is leveraging agents and LLMs without worrying too much on context/up-to-date data then Haystack could be a good option.\nIf the focus is to eventually use RAG then our framework could help.
Additionally, there might be a potential route where both are used, depending on the use case.
Feel free to dm if you want to chat further on this!","created_at":"2023-11-21T21:42:18Z","created_at_i":1700602938,"objectID":"38370678","parent_id":38370607,"story_id":38368570,"story_title":"Show HN: Neum AI \u2013 Open-source large-scale RAG framework","story_url":"https://github.com/NeumTry/NeumAI","updated_at":"2024-09-20T15:42:58Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"oryx1729"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["haystack","rag","framework"],"value":"Hey HN,
We built a live demo to compare DeepSeek R1 and OpenAI\u2019s o1 model on a simple RAG pipeline using Haystack, our open-source AI framework. Both pipelines use the same retrieval and prompts to ensure a fair comparison.
Would love to hear your thoughts -- how do open-weight models stack up against proprietary ones for RAG?
blog post + demo link: https://www.deepset.ai/blog/deepseek-openai-comparison-by-ha..."},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["rag"],"value":"Show HN: Live Demo \u2013 DeepSeek R1 vs. OpenAI O1 for RAG"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://r1-demo.deepset.ai"}},"_tags":["story","author_oryx1729","story_42879717","show_hn"],"author":"oryx1729","children":[42879744],"created_at":"2025-01-30T17:06:14Z","created_at_i":1738256774,"num_comments":0,"objectID":"42879717","points":1,"story_id":42879717,"story_text":"Hey HN,
We built a live demo to compare DeepSeek R1 and OpenAI\u2019s o1 model on a simple RAG pipeline using Haystack, our open-source AI framework. Both pipelines use the same retrieval and prompts to ensure a fair comparison.
Would love to hear your thoughts -- how do open-weight models stack up against proprietary ones for RAG?
blog post + demo link: https://www.deepset.ai/blog/deepseek-openai-comparison-by-ha...","title":"Show HN: Live Demo \u2013 DeepSeek R1 vs. OpenAI O1 for RAG","updated_at":"2025-01-30T17:11:21Z","url":"https://r1-demo.deepset.ai"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"doener"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["haystack","rag","framework"],"value":"Haystack: Open-Source AI Framework for Production Ready Agents, RAG"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["haystack"],"value":"https://haystack.deepset.ai/"}},"_tags":["story","author_doener","story_48658095"],"author":"doener","children":[48658699,48658839,48658899,48659829,48661006,48661349,48661995,48663294,48664518,48670533,48683035],"created_at":"2026-06-24T11:21:32Z","created_at_i":1782300092,"num_comments":22,"objectID":"48658095","points":90,"story_id":48658095,"title":"Haystack: Open-Source AI Framework for Production Ready Agents, RAG","updated_at":"2026-07-07T22:14:08Z","url":"https://haystack.deepset.ai/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"t-kalinowski"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["haystack","rag","framework"],"value":"One project I\u2019d add to the comparison is raghilda from Posit:
https://posit-dev.github.io/raghilda/
raghilda is deliberately narrower than LangGraph/Haystack/etc. It focuses on the RAG parts: document ingestion, chunking, embedding, storage, and retrieval. It\u2019s meant to leave the pieces visible and replaceable rather than wrap the whole application in a framework.
It also composes well with the other parts of the Posit ecosystem: chatlas for LLM interaction, and Shiny for building interactive apps around the result.
My bias is that orchestration is often better hand-rolled for the specific application. Once the framework gets too far removed from the actual steps, it tends to bring in a lot of complexity you probably don\u2019t need. In the end, most of these pieces are not that complicated.
Disclosure: I\u2019m one of the authors of raghilda."},"story_title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["haystack","rag","framework"],"value":"Haystack: Open-Source AI Framework for Production Ready Agents, RAG"},"story_url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["haystack"],"value":"https://haystack.deepset.ai/"}},"_tags":["comment","author_t-kalinowski","story_48658095"],"author":"t-kalinowski","comment_text":"One project I\u2019d add to the comparison is raghilda from Posit:
https://posit-dev.github.io/raghilda/
raghilda is deliberately narrower than LangGraph/Haystack/etc. It focuses on the RAG parts: document ingestion, chunking, embedding, storage, and retrieval. It\u2019s meant to leave the pieces visible and replaceable rather than wrap the whole application in a framework.
It also composes well with the other parts of the Posit ecosystem: chatlas for LLM interaction, and Shiny for building interactive apps around the result.
My bias is that orchestration is often better hand-rolled for the specific application. Once the framework gets too far removed from the actual steps, it tends to bring in a lot of complexity you probably don\u2019t need. In the end, most of these pieces are not that complicated.
Disclosure: I\u2019m one of the authors of raghilda.","created_at":"2026-06-24T19:34:12Z","created_at_i":1782329652,"objectID":"48664681","parent_id":48658839,"story_id":48658095,"story_title":"Haystack: Open-Source AI Framework for Production Ready Agents, RAG","story_url":"https://haystack.deepset.ai/","updated_at":"2026-06-24T19:43:50Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"elijahbenizzy"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["haystack","rag","framework"],"value":"Hey HN, In the months since we initially released Burr (https://news.ycombinator.com/item?id=39917364), we have been hard at work. We wanted to share some of the most exciting changes we\u2019ve made to build Burr out as a full-stack development framework for AI agents.
In case you don\u2019t recall, Burr is an open-source python library that makes it easier to build and debug GenAI applications & agents by representing them as graphs of simple python objects/functions. Burr only abstracts away system-level concerns (state persistence, debugging, observability), and does not dictate the way you interact with LLMs. Burr comes with a host of capabilities including an open-source UI for monitoring and observing. Burr competes with (and complements) libraries such as Haystack and LangGraph, differentiating with a focus on simpler graph state and observability constructs. We value clarity and customization over terseness (we do not have a graduation problem).
You can find the repository here: https://github.com/dagworks-inc/burr.
We are really excited about the following new features:
- Recursive, Parallel Agents: Model multi-agent hierarchies and track directly in the UI
- UI Annotations: Mark production data to review and gather post-execution evaluation/test datasets
- OpenTelemetry Integration: Log to OpenTelemetry and ingest OTel in the Burr UI to improve and customize visibility
- Reloading, Time Travel, and Forking: Debug by reloading any point in the execution history to replay and fix issues.
- Production-Ready Monitoring: Deploy with a simple self-hosted S3-based system.
Since releasing, people are building & successfully shipping: concierge agents for slack, voice answer agents for restaurants, agents over RAG systems, co-pilots for internal business workflows, to name a few. On top of this we have an exciting set of blog posts, writeups, and user testimony \u2013 we\u2019ll be sharing this + more links to get started in a comment below!"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["framework"],"value":"Show HN: Updates on Burr (OS) \u2013 a full-stack AI agent framework"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://burr.dagworks.io/"}},"_tags":["story","author_elijahbenizzy","story_42183809","show_hn"],"author":"elijahbenizzy","children":[42183880,42184291],"created_at":"2024-11-19T14:30:09Z","created_at_i":1732026609,"num_comments":1,"objectID":"42183809","points":8,"story_id":42183809,"story_text":"Hey HN, In the months since we initially released Burr (https://news.ycombinator.com/item?id=39917364), we have been hard at work. We wanted to share some of the most exciting changes we\u2019ve made to build Burr out as a full-stack development framework for AI agents.
In case you don\u2019t recall, Burr is an open-source python library that makes it easier to build and debug GenAI applications & agents by representing them as graphs of simple python objects/functions. Burr only abstracts away system-level concerns (state persistence, debugging, observability), and does not dictate the way you interact with LLMs. Burr comes with a host of capabilities including an open-source UI for monitoring and observing. Burr competes with (and complements) libraries such as Haystack and LangGraph, differentiating with a focus on simpler graph state and observability constructs. We value clarity and customization over terseness (we do not have a graduation problem).
You can find the repository here: https://github.com/dagworks-inc/burr.
We are really excited about the following new features:
- Recursive, Parallel Agents: Model multi-agent hierarchies and track directly in the UI
- UI Annotations: Mark production data to review and gather post-execution evaluation/test datasets
- OpenTelemetry Integration: Log to OpenTelemetry and ingest OTel in the Burr UI to improve and customize visibility
- Reloading, Time Travel, and Forking: Debug by reloading any point in the execution history to replay and fix issues.
- Production-Ready Monitoring: Deploy with a simple self-hosted S3-based system.
Since releasing, people are building & successfully shipping: concierge agents for slack, voice answer agents for restaurants, agents over RAG systems, co-pilots for internal business workflows, to name a few. On top of this we have an exciting set of blog posts, writeups, and user testimony \u2013 we\u2019ll be sharing this + more links to get started in a comment below!","title":"Show HN: Updates on Burr (OS) \u2013 a full-stack AI agent framework","updated_at":"2024-11-19T18:35:51Z","url":"https://burr.dagworks.io/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"varshith17"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["haystack","rag","framework"],"value":"I\u2019ve been working on a project called Valori, a Python-native vector database I built from the ground up \u2014 not by reinventing every algorithm, but by wiring together efficient, well-known indexing and search techniques into a cohesive, hackable framework.
The idea came from my frustration with existing vector DBs that were either too heavy for experimentation or too opaque to modify. I wanted something simple, modular, and extensible \u2014 so I built it.
What it does:
Lets you store, index, and search high-dimensional vectors
Supports multiple indices (Flat, HNSW, IVF, LSH, Annoy)
Has memory, disk, and hybrid storage backends
Includes a full document processing pipeline (parsing, cleaning, chunking, embedding)
Offers quantization, persistence, and plugin-based extensibility
All written in Python, integrated with NumPy, and production-tested with logging and monitoring built in.
Install:
pip install valori
GitHub: https://github.com/varshith-Git/valori
PyPI: https://pypi.org/project/valori
I\u2019d love to hear your thoughts \u2014
What\u2019s missing for you in current vector DBs?
If you\u2019ve built LLM or RAG systems, what do you wish a lightweight, pure Python DB like this handled better?
Would you prefer tighter integrations (LangChain, Haystack, etc.) or a more \u201cbuild-it-yourself\u201d style?
Feedback, criticism, or collaboration ideas are all welcome.\n\u2014 Varshith\n(varshith.gudur17@gmail.com\n)"},"title":{"matchLevel":"none","matchedWords":[],"value":"Valori \u2013 A Python-native Vector Database I built from scratch"}},"_tags":["story","author_varshith17","story_45864900","ask_hn"],"author":"varshith17","children":[45864906,45865418,45865609,45870302,45885653],"created_at":"2025-11-09T11:52:18Z","created_at_i":1762689138,"num_comments":11,"objectID":"45864900","points":9,"story_id":45864900,"story_text":"I\u2019ve been working on a project called Valori, a Python-native vector database I built from the ground up \u2014 not by reinventing every algorithm, but by wiring together efficient, well-known indexing and search techniques into a cohesive, hackable framework.
The idea came from my frustration with existing vector DBs that were either too heavy for experimentation or too opaque to modify. I wanted something simple, modular, and extensible \u2014 so I built it.
What it does:
Lets you store, index, and search high-dimensional vectors
Supports multiple indices (Flat, HNSW, IVF, LSH, Annoy)
Has memory, disk, and hybrid storage backends
Includes a full document processing pipeline (parsing, cleaning, chunking, embedding)
Offers quantization, persistence, and plugin-based extensibility
All written in Python, integrated with NumPy, and production-tested with logging and monitoring built in.
Install:
pip install valori
GitHub: https://github.com/varshith-Git/valori
PyPI: https://pypi.org/project/valori
I\u2019d love to hear your thoughts \u2014
What\u2019s missing for you in current vector DBs?
If you\u2019ve built LLM or RAG systems, what do you wish a lightweight, pure Python DB like this handled better?
Would you prefer tighter integrations (LangChain, Haystack, etc.) or a more \u201cbuild-it-yourself\u201d style?
Feedback, criticism, or collaboration ideas are all welcome.\n\u2014 Varshith\n(varshith.gudur17@gmail.com\n)","title":"Valori \u2013 A Python-native Vector Database I built from scratch","updated_at":"2026-03-05T23:00:11Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"enouri"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["haystack","rag","framework"],"value":"Location: Paris, France
Remote: Yes (for flexibility, not a requirement)
Willing to relocate: Yes
I\u2019m Nouri, a senior engineer and architect with 15+ years of experience. I\u2019ve worked for major companies like Intel, Airbus, Canon, CEA (French Atomic Agency), and SNCF, and I\u2019ve also founded and built 11 startups, mostly bootstrapped and focused on SaaS, mobile B2B, and B2C.
I\u2019m now wrapping up my latest startup (no PMF \u2014 product okay but no revenue model): Soir\u00e9es, an AI photo-sharing platform used in 13 countries by 120K+ users, powering events like Paris Fashion Weeks, Euro 2024, and the Paris Olympics 2024. I handled everything from architecture and product to growth, GTM, and AI-powered privacy.
What I\u2019m looking for: a stable, long-term role on a serious product, ideally where I can contribute across architecture, AI, product thinking, and execution. I thrive in roles that blend code, systems, and business alignment.
Technologies
Cloud: AWS (Lambda, API Gateway, DynamoDB, Fargate, SageMaker), GCP, Aliyun, OpenStack
Infra & DevOps: Kubernetes, Docker, Terraform, Serverless Framework, CI/CD
Backend: Python, Node.js, TypeScript, GraphQL, REST, gRPC
Frontend & Mobile: React, React Native, Swift, Kotlin, Fastlane
Data & Observability: Kafka, Kinesis, Airflow, Spark, ClickHouse, Redis, ELK, Prometheus
AI/ML: Haystack, LangChain, LlamaIndex, Whisper, YOLO, OpenCV, HuggingFace, ComfyUI, SageMaker
RAG & Vector DBs: Pinecone, Weaviate, Qdrant, OpenSearch (ES)
Product & GTM: SPIN Selling, PMF Discovery, AARRR, Growth Loops, Stripe Connect
R\u00e9sum\u00e9/CV: https://e-nouri.com/cv.pdf\nEmail: nouri [:at:] e-nouri [:dot:] com"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: Who wants to be hired? (July 2025)"}},"_tags":["comment","author_enouri","story_44434574"],"author":"enouri","comment_text":"Location: Paris, France
Remote: Yes (for flexibility, not a requirement)
Willing to relocate: Yes
I\u2019m Nouri, a senior engineer and architect with 15+ years of experience. I\u2019ve worked for major companies like Intel, Airbus, Canon, CEA (French Atomic Agency), and SNCF, and I\u2019ve also founded and built 11 startups, mostly bootstrapped and focused on SaaS, mobile B2B, and B2C.
I\u2019m now wrapping up my latest startup (no PMF \u2014 product okay but no revenue model): Soir\u00e9es, an AI photo-sharing platform used in 13 countries by 120K+ users, powering events like Paris Fashion Weeks, Euro 2024, and the Paris Olympics 2024. I handled everything from architecture and product to growth, GTM, and AI-powered privacy.
What I\u2019m looking for: a stable, long-term role on a serious product, ideally where I can contribute across architecture, AI, product thinking, and execution. I thrive in roles that blend code, systems, and business alignment.
Technologies
Cloud: AWS (Lambda, API Gateway, DynamoDB, Fargate, SageMaker), GCP, Aliyun, OpenStack
Infra & DevOps: Kubernetes, Docker, Terraform, Serverless Framework, CI/CD
Backend: Python, Node.js, TypeScript, GraphQL, REST, gRPC
Frontend & Mobile: React, React Native, Swift, Kotlin, Fastlane
Data & Observability: Kafka, Kinesis, Airflow, Spark, ClickHouse, Redis, ELK, Prometheus
AI/ML: Haystack, LangChain, LlamaIndex, Whisper, YOLO, OpenCV, HuggingFace, ComfyUI, SageMaker
RAG & Vector DBs: Pinecone, Weaviate, Qdrant, OpenSearch (ES)
Product & GTM: SPIN Selling, PMF Discovery, AARRR, Growth Loops, Stripe Connect
R\u00e9sum\u00e9/CV: https://e-nouri.com/cv.pdf\nEmail: nouri [:at:] e-nouri [:dot:] com","created_at":"2025-07-03T15:04:30Z","created_at_i":1751555070,"objectID":"44455799","parent_id":44434574,"story_id":44434574,"story_title":"Ask HN: Who wants to be hired? (July 2025)","updated_at":"2025-07-03T15:10:55Z"}],"hitsPerPage":20,"nbHits":7,"nbPages":1,"page":0,"params":"query=haystack+rag+framework&advancedSyntax=true&analyticsTags=backend","processingTimeMS":16,"processingTimingsMS":{"_request":{"roundTrip":23},"afterFetch":{"merge":{"total":1},"total":1},"fetch":{"query":10,"scanning":3,"total":14},"total":16},"query":"haystack rag framework","serverTimeMS":17}