Qdrant

Vector Space Wall

Submit your testimonialBuild your own wall? It's free 👉
Rishabh Gupta

Computer Scientist @ Adobe

Qdrant Thanks! We're using Qdrant as our core vector DB — fast, flexible, and awesome for PipesHub!

Vigneshwaran S

Co-Founder

Qdrant is our goto vector search for multiple scenarios

Swattik Maiti

Data Science @ American Express

It was my first time using Qdrant VectorDB and I was genuinely impressed by how seamless Qdrant was to work with—powerful, intuitive, and rock-solid performance even under fast iteration cycles. Would love to explore more ways to collaborate or contribute—Qdrant is definitely a core part of my GenAI toolkit moving forward!
It was my first time using Qdrant VectorDB and I was genuinely impressed by how seamless Qdrant was to work with—powerful, intuitive, and rock-solid performance even under fast iteration cycles. Would love to explore more ways to collaborate or contribute—Qdrant is definitely a core part of my GenAI toolkit moving forward!

Josef Kerner

AI & Data Architect

Just finished large project on qdrant and have to say that your performance and features are out of this world, great work:)
Just finished large project on qdrant and have to say that your performance and features are out of this world, great work:)

Jean-Louis Quéguiner

Founder

Qdrant is the best. Thank you for making this.

Zharko Popovski

Software Developer

Qdrant works seamlessly and super fast without any issues, regardless of the client, programming language, or framework you are using. I absolutely recommend it!
Qdrant works seamlessly and super fast without any issues, regardless of the client, programming language, or framework you are using. I absolutely recommend it!

Ehab Hussein

Principal AI Engineer

Qdrant is always my goto vectordb for any project. Keep being Awesome!

Paul Iusztin

Senior ML/AI Engineer

Found Qdrant super powerful when building my open-source courses and writing my book 🔥

Asimansu Bera

Product Solution Architect

Vector data is an emerging technology that I’m excited to explore in the coming days. It has the potential to revolutionize data processing and enhance AI-based applications. This technology offers great learning opportunities for database developers, consultants, and professionals who are eager to stay ahead. #VectorDatabase #Qdrant
Vector data is an emerging technology that I’m excited to explore in the coming days. It has the potential to revolutionize data processing and enhance AI-based applications. This technology offers great learning opportunities for database developers, consultants, and professionals who are eager to stay ahead. #VectorDatabase #Qdrant

Luis Fernando Solis Navarro

Data Scientist

I feel in a space when I use Qdrant 🥰

Rubén Quilez Domingo

Artificial Intelligence Engineer

What I truly appreciate about Qdrant is its Discovery API, which allows me to dynamically adjust the vector space in real-time. Using positive and negative vectors, I can fine-tune the relevance of each search, optimizing the process continuously. This feature significantly enhances the precision of every query. A must-have tool for RAG systems and vector search! 🚀
What I truly appreciate about Qdrant is its Discovery API, which allows me to dynamically adjust the vector space in real-time. Using positive and negative vectors, I can fine-tune the relevance of each search, optimizing the process continuously. This feature significantly enhances the precision of every query. A must-have tool for RAG systems and vector search! 🚀

Nguyen Thu

AI Engineer

Qdrant is an open-source vector search engine designed for efficiently managing and searching large-scale high-dimensional data. The key features are high performance vector search, scalability, and support complex-queries. Our product used Qdrant as vector DB for building RAG application.
Qdrant is an open-source vector search engine designed for efficiently managing and searching large-scale high-dimensional data. The key features are high performance vector search, scalability, and support complex-queries. Our product used Qdrant as vector DB for building RAG application.

Alex Dantart

Humanizing tech (especially AI)

I’m excited to share that I’ve been using Qdrant in my projects, and it’s been performing exceptionally well! The speed and efficiency of the vector search engine have really improved my workflow. I’m thrilled with the results and can’t wait to explore more possibilities with it.
I’m excited to share that I’ve been using Qdrant in my projects, and it’s been performing exceptionally well! The speed and efficiency of the vector search engine have really improved my workflow. I’m thrilled with the results and can’t wait to explore more possibilities with it.

João Campista

Advancing Generative AI

🚀 I’m leveraging Qdrant to build applications featuring smart recommendation systems and Retrieval Augmented Generation (RAG) pipelines. With its efficient handling of high-dimensional vector data and flexible query options, it’s transforming the way I develop projects by delivering incredible personalization and high performance. 🔍💡 If you're also exploring recommendation engines and want to optimize your solutions with RAG, I highly recommend checking out Qdrant. The scalability and precision potential is truly impressive! ⚡ #Qdrant #SistemasDeRecomendação #RAG #IA #MachineLearning #BuscaVetorial #InovaçãoTech #VectorStore #LLM
🚀 I’m leveraging Qdrant to build applications featuring smart recommendation systems and Retrieval Augmented Generation (RAG) pipelines. With its efficient handling of high-dimensional vector data and flexible query options, it’s transforming the way I develop projects by delivering incredible personalization and high performance. 🔍💡 If you're also exploring recommendation engines and want to optimize your solutions with RAG, I highly recommend checking out Qdrant. The scalability and precision potential is truly impressive! ⚡ #Qdrant #SistemasDeRecomendação #RAG #IA #MachineLearning #BuscaVetorial #InovaçãoTech #VectorStore #LLM

Nicolò Fretti

AI Engineer

I've implemented Qdrant alongside LlamaIndex for semantic search in several projects, and I can confidently say it's a powerful and user-friendly open-source vector search engine. Its clean API and open-source nature make it incredibly powerful yet easy to use, enabling efficient similarity searches across large datasets. Looking forward to contributing more to this growing ecosystem! Kudos to the team for creating such a robust tool! 💪 #Qdrant #VectorSearch #OpenSource
I've implemented Qdrant alongside LlamaIndex for semantic search in several projects, and I can confidently say it's a powerful and user-friendly open-source vector search engine. Its clean API and open-source nature make it incredibly powerful yet easy to use, enabling efficient similarity searches across large datasets. Looking forward to contributing more to this growing ecosystem! Kudos to the team for creating such a robust tool! 💪 #Qdrant #VectorSearch #OpenSource

Andrey Holz

Head of DS & ML

I use it for my RAG solutions. it's simple, powerful and reliable!

Vi Ngo Van

NLP/Machine Learning Team Lead

Our product used Qdrant as a vector database for building RAG (Retrieval-Augmented Generation) applications. Additionally, Qdrant's powerful similarity search capabilities make it an excellent choice for implementing sophisticated recommendation systems, enabling businesses to deliver personalized content and product suggestions to their users with high accuracy and speed.
Our product used Qdrant as a vector database for building RAG (Retrieval-Augmented Generation) applications. Additionally, Qdrant's powerful similarity search capabilities make it an excellent choice for implementing sophisticated recommendation systems, enabling businesses to deliver personalized content and product suggestions to their users with high accuracy and speed.

Rishav kumar

Machine learning

There is a lot of vector db you will come across nowadays , but qdrant is the one which is moving fast on the scale of technical advancement as well inference speed. I am going to merge it for my agentic Indian railway rag system .
There is a lot of vector db you will come across nowadays , but qdrant is the one which is moving fast on the scale of technical advancement as well inference speed. I am going to merge it for my agentic Indian railway rag system .

Alberto Dughera

BD Manager leading the AI Revolution

Having integrated Qdrant with LlamaIndex in multiple projects, I can confidently say it’s a game-changer! This open-source vector search engine is not only powerful but also incredibly user-friendly: with its sleek API, it makes similarity searches across massive datasets a breeze—delivering speed, precision, and simplicity all in one package!
Having integrated Qdrant with LlamaIndex in multiple projects, I can confidently say it’s a game-changer! This open-source vector search engine is not only powerful but also incredibly user-friendly: with its sleek API, it makes similarity searches across massive datasets a breeze—delivering speed, precision, and simplicity all in one package!

Athos G

Principal Software Engineer

I've been using Qdrant to build knowledgebase systems that require an effective, cheap and fast vector database. I'm quite impressed at the results, especially with the open source container which performs faster than even some proprietary equivalents. …more
I've been using Qdrant to build knowledgebase systems that require an effective, cheap and fast vector database. I'm quite impressed at the results, especially with the open source container which performs faster than even some proprietary equivalents. …more

Joe Chiarella

Senior Software Engineer

Qdrant is what powers AdPipe's powerful video search, and I had a blast implementing it. I can't recommend it enough.

Pavel Adámek

BE developer

We are using Qdrant for anything vector-related. It’s powering our media search alongside our semantic search. It’s a really great and impressive tool that I recommend to everyone.
We are using Qdrant for anything vector-related. It’s powering our media search alongside our semantic search. It’s a really great and impressive tool that I recommend to everyone.

Filipp Trigub

CTO

Recently started using Qdrant and am very happy with the performance!

Mohamed Jama

Principal, Technology

I think #qdrant is really one of the best vector databases out there. Kudos to the team.

Tales Marra

Machine Learning Engineer

Thank you for the work you are doing! Qdrant is awesome!

Gábor Vecsei

Lead AI | Group Lead

Setting up Qdrant for a dev project was super quick, and honestly, this convinced me to use it (when compared with other tools). Now we use it for every project.
Setting up Qdrant for a dev project was super quick, and honestly, this convinced me to use it (when compared with other tools). Now we use it for every project.

Analytics and Artificial Intelligence

Garrett Eastham

Qdrant remains my favorite and primary Vector DB of choice!

Davinder Singh

Senior Principal Engineer

I used almost all major vector DBs, but Qdrant is always standout in term of technology, performance, integration and adaptability to most of scenarios .
I used almost all major vector DBs, but Qdrant is always standout in term of technology, performance, integration and adaptability to most of scenarios .

Matt Suiche

Detection Engineering

♥️ Qdrant

Jimi Vaubien View Jimi Vaubien’s profile

🏄🏽‍♂️ Surfing the AI wave 🎥 YT @bitswired 🧑🏽‍💻bitswired.com/blog | AI Engineer

Fastembed is dope I used it in so many PoC Great work guys 🚀

Dustin Hughes

Technical Director

Keep up the good work. I have been using it for a long time, and it keeps improving with every release.

Shashi Kiran Deshetti

Co-Founder at Factly

Clean code leads to simplicity. Simplicity is one of the reasons we switched from Milvus to Qdrant at Factly Media & Research. It is working out pretty well for us so far!
Clean code leads to simplicity. Simplicity is one of the reasons we switched from Milvus to Qdrant at Factly Media & Research. It is working out pretty well for us so far!

Devanshu Mahapatra

Research

A big thank you to the Qdrant team for producing such an excellent tool! If you're looking for a dependable vector search engine, I strongly recommend giving Qdrant a try.
A big thank you to the Qdrant team for producing such an excellent tool! If you're looking for a dependable vector search engine, I strongly recommend giving Qdrant a try.

Charlie Masters

ML Research @ Xapien

Qdrant has been building great tooling for some time in the information retrieval space- I remember it being a breath of fresh air moving away from faiss flat indexes and having simple, scalable and feature filled vector indexing. I especially appreciated the work on semantic search with metadata filters- the articles on extending the HNSW algorithm contained some seriously cool stuff!
Qdrant has been building great tooling for some time in the information retrieval space- I remember it being a breath of fresh air moving away from faiss flat indexes and having simple, scalable and feature filled vector indexing. I especially appreciated the work on semantic search with metadata filters- the articles on extending the HNSW algorithm contained some seriously cool stuff!

Vincent Teyssier

AI | Strategy | Future of Wealth

Start to add many metadata fields and have more than 25Gb of index space and you will see Chroma becoming fully non functional. In our own tests Qdrant handles 80Gb indexes with 16 metadata fields on each vectors without a sweat)
Start to add many metadata fields and have more than 25Gb of index space and you will see Chroma becoming fully non functional. In our own tests Qdrant handles 80Gb indexes with 16 metadata fields on each vectors without a sweat)

Tom Kiljo

Senior Software Architect at Nitor

We have been using #Qdrant behind an easy to use (non)developer friendly custom API as part of our infra setup for internal development and learning experiences with embedded data. Looking forward to making more cool things with #Qdrant in the near future.
We have been using #Qdrant behind an easy to use (non)developer friendly custom API as part of our infra setup for internal development and learning experiences with embedded data. Looking forward to making more cool things with #Qdrant in the near future.

Arnaldo Neto

Software Developer

The unparalleled flexibility it provides in both queries and deployment instantly caught our attention. Without a moment's hesitation, we've decided to make the transition. With Qdrant leading the way, we're confident that our application will soar to unprecedented levels of success! 🚀
The unparalleled flexibility it provides in both queries and deployment instantly caught our attention. Without a moment's hesitation, we've decided to make the transition. With Qdrant leading the way, we're confident that our application will soar to unprecedented levels of success! 🚀

Shawn Rushefsky

AI for the people

Qdrant has been my go-to for RAG applications. It’s super easy to set up, and blazing fast. I’d happily replace elasticsearch/opensearch with it, too, if I was still operating either of those.
Qdrant has been my go-to for RAG applications. It’s super easy to set up, and blazing fast. I’d happily replace elasticsearch/opensearch with it, too, if I was still operating either of those.

Mohamad Fadil

Machine Learning Engineer

We are using qdrant as our vectorDB for our RAG solution. By far the best among the ones we tried in terms of performance and ease of setting up. Self hosted on our cloud server. We were able to run with full configuration in few hours.
We are using qdrant as our vectorDB for our RAG solution. By far the best among the ones we tried in terms of performance and ease of setting up. Self hosted on our cloud server. We were able to run with full configuration in few hours.

Porimol Chandro

Machine Learning Engineer (MLOps) at Roche

Three months ago, I was planning to develop an Interactive Python and SQL Learning Companion ChatBot based on CodeLlama2, but the problem I was facing performance lacking issue as I intended to use only LLM without training and without any RAG/VectorDB. Later on to improve the performance without training the model from scratch, I decided to use RAG/VectorDB and I findout that Qdrant perfectly match with my project goal. Why I used and also planning to use further? Well, -> It's intuitive and Langchain utilities available -> Easy to use -> Broader community supports -> Open Source 😎 -> and Many more
Three months ago, I was planning to develop an Interactive Python and SQL Learning Companion ChatBot based on CodeLlama2, but the problem I was facing performance lacking issue as I intended to use only LLM without training and without any RAG/VectorDB. Later on to improve the performance without training the model from scratch, I decided to use RAG/VectorDB and I findout that Qdrant perfectly match with my project goal. Why I used and also planning to use further? Well, -> It's intuitive and Langchain utilities available -> Easy to use -> Broader community supports -> Open Source 😎 -> and Many more

Christoph Michael

Zero Labelling. Full Data Intelligence.

We are currently scaling #Qdrant to use it on autonomous driving use cases. This means hundreds of millions of vectors being created, stored and used.
We are currently scaling #Qdrant to use it on autonomous driving use cases. This means hundreds of millions of vectors being created, stored and used.

Gerhard Stockinger

Digitalization 🧠 Smart Innovations through Clever Use of AI!

Your claim sounds perfect: don’t get lost in vector space. And your service is really easy to set up. Just to let you know, I’ve got the service up and running, and I’m even using it with my iPad by the pool!
Your claim sounds perfect: don’t get lost in vector space. And your service is really easy to set up. Just to let you know, I’ve got the service up and running, and I’m even using it with my iPad by the pool!

Kyle Tobin

Generative AI Products & Cognizant

We LOVE Qdrant! The exceptional engineering, strong business value, and outstanding team behind the product drove our choice. Thank you for your great contribution to the technology community!
We LOVE Qdrant! The exceptional engineering, strong business value, and outstanding team behind the product drove our choice. Thank you for your great contribution to the technology community!

Dylan Yap Jin Quan

Data scientist | NLP | LLM

Qdrant is one of the best vector databases I explored so far!

Robert Caulk

Founder @ Emergent Methods

Indeed, finally outsourcing performance and accuracy to the Qdrant team turned out to be one of the biggest catalysts for our company :)
Indeed, finally outsourcing performance and accuracy to the Qdrant team turned out to be one of the biggest catalysts for our company :)

William Lewis

Chief Technology Officer

Any day Qdrant! We are using it at Beaconcross Technologies to power our vector databases. Some of the things which we found very useful as we navigated the vector database space: 1. Easy installation (cloud/docker) 2. No performance/accuracy issues yet 3. Useful developer documentation + blogs/articles 4. Good support …see more
Any day Qdrant! We are using it at Beaconcross Technologies to power our vector databases. Some of the things which we found very useful as we navigated the vector database space: 1. Easy installation (cloud/docker) 2. No performance/accuracy issues yet 3. Useful developer documentation + blogs/articles 4. Good support …see more

Tim Kolberg

CTO | Tech Strategist

The Roadmap for Qdrant in 2024 looks fantastic 🚀 Everyone working with #vectorsearch, #ai, #llms should be aware of those guys.
The Roadmap for Qdrant in 2024 looks fantastic 🚀 Everyone working with #vectorsearch, #ai, #llms should be aware of those guys.

Nicola Procopio

Senior Data Scientist

I recommend Qdrant, easy to install and use, great documentation and community, a lot of features.

Xavier Portilla Edo

Infrastructure Team Lead at Voiceflow

🌲 Farewell Pinecone, Hello Qdrant: As part of our commitment to delivering cutting-edge solutions, we've bid adieu to Pinecone as a vector database and embraced the powerful capabilities of QDrant for our knowledge base. This strategic move enhances our system's efficiency and takes us a step closer to achieving excellence in Conversational AI technology. 🚀💡
🌲 Farewell Pinecone, Hello Qdrant: As part of our commitment to delivering cutting-edge solutions, we've bid adieu to Pinecone as a vector database and embraced the powerful capabilities of QDrant for our knowledge base. This strategic move enhances our system's efficiency and takes us a step closer to achieving excellence in Conversational AI technology. 🚀💡

Vincent Teyssier

AI | Strategy

Why we chose Qdrant as our vector database? Performance is a key element of LLM applications. Though it is not a bottleneck while prototyping, it becomes one when scaling. As many, we started our journey with the usual open source suspect where it comes to vector databases, but on our scaling journey, we quickly realised that performance optimisations were needed. We started digging into the code to add indexes and other tweaks but it was hopeless. Some libraries are just not mature enough to handle production workloads. When looking at more robust alternatives, we found a few, but ultimately settled with Qdrant. Why? 📈 Firstly, performance. It handled indexing 1.2M vectors with 16 metadata fields for each vector without a sweat and with no performance degradation. Similarity queries or scrolls run in less than 0.3s. So we decided to double that to 2.4M vectors, and it's as is we were inserting our first vector! 📄 Secondly, memory efficiency. Our final database storage folder is 8x smaller than on our previous database. This allows us to put the whole dataset in smaller instances, hence saving significant infrastructure costs. We now even deploy the full dataset in dev environment. 📦 Thirdly, the embedded capabilities. Beyond simple search and similarity, it hosts a bunch of very nice features around recommendation engines, adding positive and negative examples for better spacial narrowing, efficient multi-tenancy, and many more.... 🤼 Finally, the Qdrant team. Andre Zayarni and his team are doing a fantastic job. You can feel the passion for data engineering and the strong experience leading to a robust product. Qdrant is not a marketing gimmick or a VC funds chaser, it is a real product of engineering perfectly adapted to the new production use cases that GenAI is bringing to the world. They support open source and spare no effort in helping newbies or veterans in their discord forum (thanks Tim Visée for your support).
Why we chose Qdrant as our vector database? Performance is a key element of LLM applications. Though it is not a bottleneck while prototyping, it becomes one when scaling. As many, we started our journey with the usual open source suspect where it comes to vector databases, but on our scaling journey, we quickly realised that performance optimisations were needed. We started digging into the code to add indexes and other tweaks but it was hopeless. Some libraries are just not mature enough to handle production workloads. When looking at more robust alternatives, we found a few, but ultimately settled with Qdrant. Why? 📈 Firstly, performance. It handled indexing 1.2M vectors with 16 metadata fields for each vector without a sweat and with no performance degradation. Similarity queries or scrolls run in less than 0.3s. So we decided to double that to 2.4M vectors, and it's as is we were inserting our first vector! 📄 Secondly, memory efficiency. Our final database storage folder is 8x smaller than on our previous database. This allows us to put the whole dataset in smaller instances, hence saving significant infrastructure costs. We now even deploy the full dataset in dev environment. 📦 Thirdly, the embedded capabilities. Beyond simple search and similarity, it hosts a bunch of very nice features around recommendation engines, adding positive and negative examples for better spacial narrowing, efficient multi-tenancy, and many more.... 🤼 Finally, the Qdrant team. Andre Zayarni and his team are doing a fantastic job. You can feel the passion for data engineering and the strong experience leading to a robust product. Qdrant is not a marketing gimmick or a VC funds chaser, it is a real product of engineering perfectly adapted to the new production use cases that GenAI is bringing to the world. They support open source and spare no effort in helping newbies or veterans in their discord forum (thanks Tim Visée for your support).

Sergii Shcherbak

CTO | Software Developer | Lawyer

Reliable and easy to use.

Jonathan Eisenzopf

Chief Strategy and Research Officer at Talkmap

Written in Rust, my language of choice and I think the right language long term for fast and scalable LLM applications.

Angel Luis Almaraz Sánchez

Full Stack | DevOps

Thank you, great work, Qdrant is my favorite option for similarity search.

Yasin Salimibeni View Yasin Salimibeni’s profile

AI Evangelist | Generative AI Product Designer | Entrepreneur | Mentor

Great work. I just started testing Qdrant Azure and I was impressed by the efficiency and speed. Being deploy-ready on large cloud providers is a great plus. Way to go!
Great work. I just started testing Qdrant Azure and I was impressed by the efficiency and speed. Being deploy-ready on large cloud providers is a great plus. Way to go!

Stanislas Polu

Software Engineer & Co-Founder, Dust

Qdrant's the best. By. Far.

Sivesh Sukumar

Investor at Balderton

We're using Qdrant to help segment and source Europe's next wave of extraordinary companies!

Megaklis Vasilakis

CEO at Squaredev

How I use Qdrant? Easy 😁 I only have nice words to say about your product Andre Zayarni. Retrieval augmentation is the main thing we use Qdrant for our LLM solutions. The API interface and SDK make things very easy I have to say. The in-memory feature is a competitive advantage against all other vector DBs. Scaling is easy. What else is there to say? One more thing I am using it for is in another side project that is an LLMOps solution that is currently in the builds. Storing all chat messages, for cache and analytics is super easy in Qdrant. What makes it ideal for this kind of features is the payload feature that is so nicely implemented.
How I use Qdrant? Easy 😁 I only have nice words to say about your product Andre Zayarni. Retrieval augmentation is the main thing we use Qdrant for our LLM solutions. The API interface and SDK make things very easy I have to say. The in-memory feature is a competitive advantage against all other vector DBs. Scaling is easy. What else is there to say? One more thing I am using it for is in another side project that is an LLMOps solution that is currently in the builds. Storing all chat messages, for cache and analytics is super easy in Qdrant. What makes it ideal for this kind of features is the payload feature that is so nicely implemented.

Leonard Püttmann

data scientist

Amidst the hype around vector databases, Qdrant is by far my favorite one. It's super fast (written in Rust) and open-source! At Kern AI we use Qdrant for fast document retrieval and to do quick similarity search for text data.
Amidst the hype around vector databases, Qdrant is by far my favorite one. It's super fast (written in Rust) and open-source! At Kern AI we use Qdrant for fast document retrieval and to do quick similarity search for text data.

Bayangmbe Mounmo

Text Similarity

Qdrant is just the fast and most accurate vector space database, I discovered after running their #benchmark. At DeepFile, we are using Qdrant for storing chunks of texts to enable semantic search using the #hnsw method. Stay stunned, DeepFile is preparing something for people.
Qdrant is just the fast and most accurate vector space database, I discovered after running their #benchmark. At DeepFile, we are using Qdrant for storing chunks of texts to enable semantic search using the #hnsw method. Stay stunned, DeepFile is preparing something for people.

Andrew Rove

Principal Software Engineer

We have been using Qdrant in production now for over 6 months to store vectors for cosine similarity search and it is way more stable and faster than our old ElasticSearch vector index. No merging segments, no red indexes at random times. It just works and was super easy to deploy via docker to our cluster. It’s faster, cheaper to host, and more stable, and open source to boot!
We have been using Qdrant in production now for over 6 months to store vectors for cosine similarity search and it is way more stable and faster than our old ElasticSearch vector index. No merging segments, no red indexes at random times. It just works and was super easy to deploy via docker to our cluster. It’s faster, cheaper to host, and more stable, and open source to boot!

Saksham Gupta

AI Governance Machine Learning Engineer

Looking forward to using Qdrant vector similarity search in the clinical trial space! OpenAI Embeddings + Qdrant = Match made in heaven!
Looking forward to using Qdrant vector similarity search in the clinical trial space! OpenAI Embeddings + Qdrant = Match made in heaven!

Shubham Krishna

ML Engineer @ ML6

Go ahead and checkout Qdrant. I plan to build a movie retrieval search where you can ask anything regarding a movie based on the vector embeddings generated by a LLM. It can also be used for getting recommendations.
Go ahead and checkout Qdrant. I plan to build a movie retrieval search where you can ask anything regarding a movie based on the vector embeddings generated by a LLM. It can also be used for getting recommendations.

Marcel Coetzee

Data and AI Plumber

Using Qdrant as a blazing fact vector store for a stealth project of mine. It offers fantasic functionality for semantic search ✨
Using Qdrant as a blazing fact vector store for a stealth project of mine. It offers fantasic functionality for semantic search ✨

Kwok Hing LEON

Data Science

Check out qdrant for improving searches. Bye to non-semantic KM engines.

Ken Tsui

AI and Machine Learning Leadership

I use Qdrant to build an information retrieval system, which could be used for: - open domain question answering - retrieval augmented large language model - factchecking github link: lnkd.in/gnNnAPuA
I use Qdrant to build an information retrieval system, which could be used for: - open domain question answering - retrieval augmented large language model - factchecking github link: lnkd.in/gnNnAPuA

Josh Lloyd

ML Engineer

I'm using Qdrant to search through thousands of documents to find similar text phrases for question answering. Qdrant's awesome filtering allows me to slice along metadata while I'm at it! 🚀 and it's fast ⏩🔥
I'm using Qdrant to search through thousands of documents to find similar text phrases for question answering. Qdrant's awesome filtering allows me to slice along metadata while I'm at it! 🚀 and it's fast ⏩🔥

Ankur S

Building

Quadrant is a great vector database. There is a real sense of thought behind the api!

Kenn Ejima

Hacker @ Dot Product

Just soft-launched our product that uses Qdrant to store chat with data embeddings. Previously considered Pinecone and Weaviate but finally settled with Qdrant which feels just right from a developer experience standpoint.
Just soft-launched our product that uses Qdrant to store chat with data embeddings. Previously considered Pinecone and Weaviate but finally settled with Qdrant which feels just right from a developer experience standpoint.

Paweł Saniewski

Data Engineer

Qdrant is one of the best (if not the best) vector databases out there. Currently testing out the similarity search capabilities for adding domain knowledge context to LLM prompts.
Qdrant is one of the best (if not the best) vector databases out there. Currently testing out the similarity search capabilities for adding domain knowledge context to LLM prompts.

Ikram Ali

ML Engineer

🔍 Leveraging Qdrant Vector Database for Enhanced Restaurant Similarity and Recommendation 🍽️🚀 I am thrilled to share a groundbreaking development in our quest to deliver top-notch restaurant recommendations! 🌟 We have implemented the Qdrant Vector Database in our system, revolutionizing the way we understand and analyze restaurant data. 🗺️ Qdrant Vector Database has proven to be an invaluable tool in our pursuit of providing personalized dining experiences to our users. By leveraging the power of vectors, we can now capture and compare various aspects of restaurants, enabling us to uncover hidden patterns and similarities. 🧐✨ How does it work? 🤔 Qdrant Vector Database allows us to represent restaurants as vectors, where each dimension corresponds to a specific feature, such as cuisine type, ambiance, price range, and customer reviews. By organizing and indexing these vectors efficiently, we can perform lightning-fast similarity searches and generate highly accurate recommendations tailored to each user's preferences. 📊🔎 Here's what sets Qdrant Vector Database apart: 🚀 Lightning-fast: Qdrant's advanced indexing and query optimization techniques ensure that we can process vast amounts of restaurant data in real-time, delivering instant results to our users. 🔍 Precise Similarity Analysis: With Qdrant, we can identify subtle similarities and relationships between restaurants that were previously undetectable. This allows us to make intelligent recommendations based on nuanced factors that truly align with individual tastes. ⚡ Scalability and Flexibility: As our database grows, Qdrant seamlessly scales with us, effortlessly accommodating the ever-expanding world of restaurants and evolving user preferences. 🙌 we believe that personalization and quality are key to delivering exceptional experiences. With Qdrant Vector Database, we are taking our restaurant recommendations to new heights, ensuring that every user discovers their perfect dining spot effortlessly. 🌟
🔍 Leveraging Qdrant Vector Database for Enhanced Restaurant Similarity and Recommendation 🍽️🚀 I am thrilled to share a groundbreaking development in our quest to deliver top-notch restaurant recommendations! 🌟 We have implemented the Qdrant Vector Database in our system, revolutionizing the way we understand and analyze restaurant data. 🗺️ Qdrant Vector Database has proven to be an invaluable tool in our pursuit of providing personalized dining experiences to our users. By leveraging the power of vectors, we can now capture and compare various aspects of restaurants, enabling us to uncover hidden patterns and similarities. 🧐✨ How does it work? 🤔 Qdrant Vector Database allows us to represent restaurants as vectors, where each dimension corresponds to a specific feature, such as cuisine type, ambiance, price range, and customer reviews. By organizing and indexing these vectors efficiently, we can perform lightning-fast similarity searches and generate highly accurate recommendations tailored to each user's preferences. 📊🔎 Here's what sets Qdrant Vector Database apart: 🚀 Lightning-fast: Qdrant's advanced indexing and query optimization techniques ensure that we can process vast amounts of restaurant data in real-time, delivering instant results to our users. 🔍 Precise Similarity Analysis: With Qdrant, we can identify subtle similarities and relationships between restaurants that were previously undetectable. This allows us to make intelligent recommendations based on nuanced factors that truly align with individual tastes. ⚡ Scalability and Flexibility: As our database grows, Qdrant seamlessly scales with us, effortlessly accommodating the ever-expanding world of restaurants and evolving user preferences. 🙌 we believe that personalization and quality are key to delivering exceptional experiences. With Qdrant Vector Database, we are taking our restaurant recommendations to new heights, ensuring that every user discovers their perfect dining spot effortlessly. 🌟

Matteo Mistri

Data Scientist

Starting to use Qdrant for a LLM integration on very long documents. Love the possibility to have a free Cloud Cluster for free testing in this early stage.
Starting to use Qdrant for a LLM integration on very long documents. Love the possibility to have a free Cloud Cluster for free testing in this early stage.

Marcelo Wiermann

Head of Engineering

We're using Qdrant to build a few internal projects at the moment. The experience has been great so far!

Sanjay Adlakha

Director - Technology Solutions

Great stuff. Also create DML and DDL statements based on natural language user prompts using generative AI on the new Qdrant UI. Vector visualization looks slick.
Great stuff. Also create DML and DDL statements based on natural language user prompts using generative AI on the new Qdrant UI. Vector visualization looks slick.

Rishav Dash

Data Scientist

awesome stuff 🔥

Ruban Phukan View Ruban Phukan’s profile

Co-Founder & CEO

This is a very interesting approach. Great work.

Anubhav

ML and NLP Engineer

Brilliant guys!

Dhruv Gupta

Machine Learning Lead

Amazing ! 💯

Jayansh Bhartiya

Decision Scientist

This is great. Now this is a big differentiator!

Bhavish Pahwa

AI Enginner

Pretty great 🤩, congrats on achieving this!

Arya Akhare

Data Scientist, Data Analyst, ML & AI Engineer

This fast semantic search solution by Qdrant is truly impressive! Congrats on conquering the challenges of search-as-you-type scenario with clever optimization tricks. Can't wait to try it out!
This fast semantic search solution by Qdrant is truly impressive! Congrats on conquering the challenges of search-as-you-type scenario with clever optimization tricks. Can't wait to try it out!

Prithivi Da

Building a Memory Assistant for Philomaths.

Holysh*t this is huge and not many would get the implementation nuance and economics of this solution. Bravo Qdrant!

Stephen M. Walker II

CEO @ Klu.ai | Ship AI Today

For the Vector DB nerds out there, we also benchmarked their new document retrieval system on 500k documents. Averaged 429 document inserts per second and a 378ms search latency. Qdrant you're still #1 in lowest latency.
For the Vector DB nerds out there, we also benchmarked their new document retrieval system on 500k documents. Averaged 429 document inserts per second and a 378ms search latency. Qdrant you're still #1 in lowest latency.

Alireza Behtash

Co-Founder & CTO

Shoutout to Qdrant team for an amazing job with Qdrant. I recently migrated a project from pgvector into their open source Qdrant API and their batch search functionality by far beats any other tool out there in terms of speed even for plain brute force search -- no quantization and probing.
Shoutout to Qdrant team for an amazing job with Qdrant. I recently migrated a project from pgvector into their open source Qdrant API and their batch search functionality by far beats any other tool out there in terms of speed even for plain brute force search -- no quantization and probing.

Hooman Sedghamiz

Director AI and ML

Vectorstores are definitely here to stay, the objects in the world around us from image, sound, video and text become easily universal and searchable thanks to the embedding models and vectorstores. You won't be able to build a performant product using Langchain/Llama-index for many reasons such as specialized requirements for advanced filtering, hybrid search and latency optimizations. I personally recommend Qdrant. we have been using it for awhile and couldn't be happier. The earlier you start embedding your organization's data into a vectorstore the faster you open the door to numerous business opportunities and easier value generation.
Vectorstores are definitely here to stay, the objects in the world around us from image, sound, video and text become easily universal and searchable thanks to the embedding models and vectorstores. You won't be able to build a performant product using Langchain/Llama-index for many reasons such as specialized requirements for advanced filtering, hybrid search and latency optimizations. I personally recommend Qdrant. we have been using it for awhile and couldn't be happier. The earlier you start embedding your organization's data into a vectorstore the faster you open the door to numerous business opportunities and easier value generation.

Somnath Banerjee

Partner & Chief Data Scientist

We recently had to pick a vector database for our customer finding product. Vector db is a hot area with significant VC investments. However, it is also a crowded space and standing out in this crowded field—whether in terms of features or speed—is tough. Here are a few simple criteria we applied in deciding which vector db to use. 1️⃣ Deployment options: Should offer both open-source and cloud options. This lets us experiment locally and scale up on cloud as needed. Milvus, Weaviate, Qdrant etc. meet this requirement. 2️⃣ Easy to start: Their website should have getting started, tutorials, etc. documents. Qdrant has a Quick Start Guide and tutorials at the center of their home page. On the contrary, we found the Redis documentation on getting started with the vector db to be confusing. 3️⃣ Support: Quick and well organized support is essential. Qdrant has a well organized discord with a dedicated channel for “questions” and their CTO, Andrey Vasnetsov is responding to questions round the clock. On the contrary it is hard to navigate Langchain’s discord with so many channels 4️⃣ Features: Ensure filtering, index swap, token search capabilities in addition to vector search. Mostly all of them support those but good to confirm We ended up with Qdrant which met all the above criteria and so far very happy with the features, performance and support. You’re likely to get locked in to the vector db you choose. Langchain / Llamaindex integrations generally support basic querying and don't really make switching vector db easier. You’ll need custom codes to do indexing, filtering etc. So it's worth spending some time selecting  a good one to begin with. Would love to hear if any of you have more thoughts, experience with vector DB.
We recently had to pick a vector database for our customer finding product. Vector db is a hot area with significant VC investments. However, it is also a crowded space and standing out in this crowded field—whether in terms of features or speed—is tough. Here are a few simple criteria we applied in deciding which vector db to use. 1️⃣ Deployment options: Should offer both open-source and cloud options. This lets us experiment locally and scale up on cloud as needed. Milvus, Weaviate, Qdrant etc. meet this requirement. 2️⃣ Easy to start: Their website should have getting started, tutorials, etc. documents. Qdrant has a Quick Start Guide and tutorials at the center of their home page. On the contrary, we found the Redis documentation on getting started with the vector db to be confusing. 3️⃣ Support: Quick and well organized support is essential. Qdrant has a well organized discord with a dedicated channel for “questions” and their CTO, Andrey Vasnetsov is responding to questions round the clock. On the contrary it is hard to navigate Langchain’s discord with so many channels 4️⃣ Features: Ensure filtering, index swap, token search capabilities in addition to vector search. Mostly all of them support those but good to confirm We ended up with Qdrant which met all the above criteria and so far very happy with the features, performance and support. You’re likely to get locked in to the vector db you choose. Langchain / Llamaindex integrations generally support basic querying and don't really make switching vector db easier. You’ll need custom codes to do indexing, filtering etc. So it's worth spending some time selecting  a good one to begin with. Would love to hear if any of you have more thoughts, experience with vector DB.

Prabhu Raghav

CTO | CEO | Co-Founder

Filter option with HNSW is a powerful combination without a tradeoff between accuracy and latency!

Ankit Shrivastava

Senior Data Scientist at SAP

"Implementing Vector Database using Qdrant" I recently had the opportunity to experiment with Qdrant, a high-performance vector database. I wanted to understand how it functions and, more importantly, how straightforward it is to implement. The experience was absolutely exhilarating! 💥 The documentation for Qdrant is comprehensive and excellent, making it accessible for anyone to get started. A big shoutout to the team, for the recommendation. 👍
"Implementing Vector Database using Qdrant" I recently had the opportunity to experiment with Qdrant, a high-performance vector database. I wanted to understand how it functions and, more importantly, how straightforward it is to implement. The experience was absolutely exhilarating! 💥 The documentation for Qdrant is comprehensive and excellent, making it accessible for anyone to get started. A big shoutout to the team, for the recommendation. 👍

Ankit Shrivastava

Senior Data Scientist at SAP

Let's dive in and explore the power of semantic search with Qdrant! 🚀

Shraey Bhatia

Co-Founder || LLM & NLP Scientist

I quite like Qdrant. Their local docker version is handy for initial stages, and when scaling up, a cloud version is available. Plus, quite like their intuitive metadata filtering. Vector DBs also play a role for connecting recommender systems effectively.
I quite like Qdrant. Their local docker version is handy for initial stages, and when scaling up, a cloud version is available. Plus, quite like their intuitive metadata filtering. Vector DBs also play a role for connecting recommender systems effectively.

Derek Tu

Connect external data to your LLMs.

When we first started Carbon, we evaluated multiple vector database providers. In the end, we chose Qdrant, and it turned out to be one of the best decisions we made.♟️ Qdrant stood out across multiple vectors (no pun intended): • Flexibility: Qdrant offers both cloud and self-hosting options, allowing Carbon to easily start on the cloud and migrate to our own infrastructure if needed. • Simplicity: Setting up, updating, and scaling a cluster in Qdrant's hosted cloud is incredibly easy. We were able to deploy Qdrant to production in less than a day. • White-Glove Support: Andre Zayarni and his team are highly responsive on Discord. When we encountered an issue while upgrading Qdrant last week, we reached out to their team on Discord and received a resolution within minutes. Choosing the right partners can make or break your business, and I'm glad we selected Qdrant! 🚀
When we first started Carbon, we evaluated multiple vector database providers. In the end, we chose Qdrant, and it turned out to be one of the best decisions we made.♟️ Qdrant stood out across multiple vectors (no pun intended): • Flexibility: Qdrant offers both cloud and self-hosting options, allowing Carbon to easily start on the cloud and migrate to our own infrastructure if needed. • Simplicity: Setting up, updating, and scaling a cluster in Qdrant's hosted cloud is incredibly easy. We were able to deploy Qdrant to production in less than a day. • White-Glove Support: Andre Zayarni and his team are highly responsive on Discord. When we encountered an issue while upgrading Qdrant last week, we reached out to their team on Discord and received a resolution within minutes. Choosing the right partners can make or break your business, and I'm glad we selected Qdrant! 🚀

Zach Efrati

Generative AI Expert

🎆🎆 🎆 This is incredible and a great boost for many hashtag #GenAI solutions out there that leverage hashtag #VectorSearch to power contextual interactions.
🎆🎆 🎆 This is incredible and a great boost for many hashtag #GenAI solutions out there that leverage hashtag #VectorSearch to power contextual interactions.

Chris Alexiuk

🚀 Founding MLE @ Ox | 🪄 LLM Wizard

Jorge Alcantara Barroso has a great post about some great options for Indexes that you can use to power your RAG applications! If you're just delving into the Vector Database world, this should help provide insights that aid in choosing which is best for you! If you want a quick option for which you should choose for the best "out of the box" experience with the lowest time-to-implement: Choose Qdrant.
Jorge Alcantara Barroso has a great post about some great options for Indexes that you can use to power your RAG applications! If you're just delving into the Vector Database world, this should help provide insights that aid in choosing which is best for you! If you want a quick option for which you should choose for the best "out of the box" experience with the lowest time-to-implement: Choose Qdrant.

Mohamad Fadil

Software and Machine Learning Engineer

We just moved from other vectordb to Qdrant self hosted on our infrastructure. I must say i am impressed with the performance and the great experience of setting it up on our own infrastructure.
We just moved from other vectordb to Qdrant self hosted on our infrastructure. I must say i am impressed with the performance and the great experience of setting it up on our own infrastructure.

Anh Le

Lead Software Engineer

Sure using qdrant helps us bring near real-time listing recommendations in real estate industry to production!

Lev Leontiev

Enginnering Manager

Best open-source vector databases First of all , what is a vector database? Vector databases store high-dimensional data, in a lower-dimensional space while preserving important relationships and characteristics of the original data. These representations are designed to capture semantic meaning and similarities between data points. Vector databases are commonly used in various applications, including natural language processing (NLP), image processing, recommendation systems, and more. 1. Qdrant ( lnkd.in/eBfAnT8q ) (hi Andre Zayarni!) Qdrant - High-performance, massive-scale Vector Database for the next generation of AI. Also available in the cloud
Best open-source vector databases First of all , what is a vector database? Vector databases store high-dimensional data, in a lower-dimensional space while preserving important relationships and characteristics of the original data. These representations are designed to capture semantic meaning and similarities between data points. Vector databases are commonly used in various applications, including natural language processing (NLP), image processing, recommendation systems, and more. 1. Qdrant ( lnkd.in/eBfAnT8q ) (hi Andre Zayarni!) Qdrant - High-performance, massive-scale Vector Database for the next generation of AI. Also available in the cloud

Anirudh Rowjee

Databases, and Cloud Enthusiast

Thank you so much! Using io_uring in prod must be a very interesting experience, being at the bleeding edge! Thank you for the note on one-stage filtering. Also - QDrant was my intro to Vector Databases (the CMUDB Talk by Andrey Vasnetsov was a great introduction), and y'all are up to some really cool stuff.
Thank you so much! Using io_uring in prod must be a very interesting experience, being at the bleeding edge! Thank you for the note on one-stage filtering. Also - QDrant was my intro to Vector Databases (the CMUDB Talk by Andrey Vasnetsov was a great introduction), and y'all are up to some really cool stuff.

Francesco Saverio Zuppichini

Machine Learning Engineer

Whoever build the Qdrant python client deserve a medal! 🎖 Best dev experience ever so far!
Whoever build the Qdrant python client deserve a medal! 🎖 Best dev experience ever so far!

Justin Miller

Principal Software Engineer

It really is fantastic, especially over gRPC. I’m able to insert 1800 or so points a second with a single client, which is more than enough for my needs.
It really is fantastic, especially over gRPC. I’m able to insert 1800 or so points a second with a single client, which is more than enough for my needs.

Martin Dulák

Software Developer / Solution Architect

What we did (in less than 24 hours 💪) was to write a #Jira plugin that works on your issues and is able to tell you "what the truth is" and reference the source of its answer. It's written in SvelteJS, backed by Azure Functions, Semantic Kernel, OpenAI embeddings, and Qdrant vector DB (kudos for having the free managed cloud tier 👏).
What we did (in less than 24 hours 💪) was to write a #Jira plugin that works on your issues and is able to tell you "what the truth is" and reference the source of its answer. It's written in SvelteJS, backed by Azure Functions, Semantic Kernel, OpenAI embeddings, and Qdrant vector DB (kudos for having the free managed cloud tier 👏).

Ismael Garcia Carpena

Software Developer

Neural search, opensource, self-hosted, rust,... You can create a semantic search engine with your own data with very little effort using Qdrant.
Neural search, opensource, self-hosted, rust,... You can create a semantic search engine with your own data with very little effort using Qdrant.