Updated

Qdrant

Open-source vector search engine written in Rust for storing and querying high-dimensional embeddings.

18 makers' products we track use Qdrant. For vector database it's 2nd most chosen of 7 options: 18 of the 107 makers' products we know the vector database of (17%). 94 open-source projects also declare it in their code. What makers like most: It is fast and scales with strong price-performance, helped by its Rust core. Their most common complaint: As a separate server it is slower than in-process stores for small local datasets. On GitHub, 7 public pull requests moved projects to Qdrant from tools we list, and 9 moved them away.

Open sourceSelf-hostableFree tierPaid from Usage-based, no minimum

Profile checked

Works with AI agents:llms.txt
Ask your AI about this, with this page as the source:ChatGPT ↗Claude ↗Perplexity ↗

How makers use Qdrant

Makers' products we track use Qdrant for vector database (18), most often alongside OpenAI API (44%), Next.js (28%).

What Qdrant is used for

Vector Database · Open-source vector store →Heavy metadata filtering alongside vector search, self-hosted from one binary or on its managed cloud.Trade-off: One more service to deploy and keep in sync with your main database.
Pinecone53Qdrant18Weaviate16Chroma12turbopuffer8Makers' products on each, for vector database

Qdrant vs the alternatives makers chose

For vector database, 18 of the 107 makers' products we know the choice of picked Qdrant (17%), behind Pinecone (50%).

OptionMakers' productsFree forAt 10,000 usersOpen source
Pinecone53 50%~3,300 users$113/mo · StandardNoCompare →
Qdrant18 17%~1,000 users$103/mo · Standard (3 nodes, 0.5 vCPU / 4 GiB each)Yes
Weaviate16 15%~1,000 users$253/mo · FlexYesCompare →
Chroma12 11%~760 users$51/mo · StarterYesCompare →
turbopuffer8 7%None$21/mo · LaunchNoCompare →
LanceDB3 3%——YesCompare →

For vector database: how many of the 107 makers' products we know the vector database of chose each option, and what each costs at 10,000 users.

Costs are the cheapest plan that fits, from each vendor's pricing page, at the usage our stack planner assumes for a product with that many users. Free for: the users its cheapest plan a business can use stays free for.

Often paired with

Share of the 18 makers' products on Qdrant that show this choice.

Markets that build with it most

Product Hunt markets where Qdrant is in far more of the stacks we know than across every market (0%).

When to use Qdrant, and when not

Pick Qdrant if

  • You need an open-source store with heavy filtering or hybrid search, self-hosted or managed. Vector Database →

Something else fits better if

  • You already use Postgres and have up to a few million vectors: pgvector.
  • You're prototyping retrieval on your laptop, or want vectors in files with no server: Chroma.
  • You want a hosted index with nothing to operate: Pinecone or turbopuffer.
From each decision's guide, written from what makers' products chose and why.

Qdrant is a vector search engine. You store points (an ID, one or more vectors and a JSON payload) in collections, then ask for the points nearest to a query vector, optionally narrowed by conditions on the payload. It typically backs semantic search and RAG.

More about Qdrant

It runs as its own server that your app calls over REST or gRPC, with official clients for Python, JavaScript/TypeScript, Rust, Go, .NET and Java. You add a payload index on each field you filter by; these extend the HNSW graph, so filters apply during the graph search itself. A point can hold several named vectors, including sparse ones for keyword matching, so hybrid queries stay in one collection. Embeddings come from your own code, the FastEmbed library, or Cloud Inference on Qdrant Cloud.

Useful for a small team: vectors and the index can be moved from RAM to disk as data grows; many tenants can share one collection keyed by a tenant field; API keys can be read-only or scoped to single collections; and a migration tool streams data in from other vector databases. It runs from the open-source Docker image, on Qdrant Managed Cloud (AWS, GCP, Azure), or as managed clusters on your own infrastructure.

Limits: a self-hosted instance has no authentication until you configure it, and the open-source build leaves replication, backups and monitoring to you. Upgrades must go one minor version at a time and cannot be rolled back. The free cloud cluster is a single small node.

How Qdrant itself is built

3 tools, from its own code, website and Product Hunt page.

Qdrant pros and cons

What makers like most: It is fast and scales with strong price-performance, helped by its Rust core. Their most common complaint: As a separate server it is slower than in-process stores for small local datasets.

Pros

  • It is fast and scales with strong price-performance, helped by its Rust core. PH
  • Payload filtering and hybrid semantic plus boolean search work together. PH
  • Runs easily self-hosted in Docker, a common pick for local RAG and agent memory. PHHNHN 2HN 3
  • Documentation is strong, and FastEmbed provides local embeddings without an external API. PHHN
Product Hunt review tags
fast performance7
semantic search6
excellent documentation3
efficient similarity search2
scalable2
self-hosting2

Cons and watch-outs

  • As a separate server it is slower than in-process stores for small local datasets. HN
  • Setting up and operating a vector database is overkill for teams that just need working search. HNHN 2

Themes that recur in makers' words and Hacker News comments, each linked to what it summarises, with how Product Hunt tags its reviews.

5.0★ from 23 reviews on Product Hunt ↗

Who switches

More projects move away from Qdrant: 7 pull requests in, 9 out; the most come from Chroma and leave for pgvector.

Moving to QdrantMoving away
Qdrant
Chroma → Qdrant4 PRs
Pinecone → Qdrant3 PRs
Qdrant → pgvector9 PRs

Public pull requests on GitHub since Oct 2024 whose title says "X to Y" — real code changes moving a project from one tool to another, by developers in general.

Who uses it

18 makers' products, each linked to the source that shows it, and 94 open-source projects that declare it in their code. The best known on Product Hunt: Lyzr (1,312 followers), Trag (1,292 followers), Superflex (1,006 followers), April (726 followers), Conva.AI (667 followers).

The maker says so 18Declared in code 94How evidence is collected →
Cognee
CogneeOpen-source memory engine for agents that combines graph, vector and relational retrieva…

“Qdrant is easy to use, fastembed is our favorite tool and developer documentation is no match.”

Vector DatabaseMaker says so · source ↗
Byterover
ByteroverFile-based memory for agents with >92% retrieval accuracy

“Vector database enabling fast semantic search and retrieval of coding memories.”

Vector DatabaseMaker says so · source ↗
VoltAgent
VoltAgentBuild TS AI agents with n8n-style observability

“Qdrant’s performance and cloud inference options are solid. It feels reliable for production workloads across modalities.”

Vector DatabaseMaker says so · source ↗
WiseWorld
WiseWorldThe only product that turns soft skills into actionable data

“Amazing infrastructure for creating complex RAG system. Thank you”

Vector DatabaseMaker says so · source ↗
April
AprilVoice assistant app for iPhone that manages email and calendar by spoken commands, such…

“Simple to setup and start using”

Vector DatabaseMaker says so · source ↗
AICamp
AICampAI platform for businesses; Private RAG w/ 100+ LLMs

“Thanks to Qdrant, we utilize it as a vector database to store our knowledge base and uploaded file data. Our RAG would not be possible without it.”

Vector DatabaseMaker says so · source ↗
Conva.AI
Conva.AIBuild an AI Assistant for your App in 1 click

“After evaluating a bunch of Vector DBs to be our internal vector DB, we finally closed on QDrant because it was the one that scaled the best and had the best price performance ratio”

Vector DatabaseMaker says so · source ↗
Lyzr
LyzrPrivate Agent SDKs to build Generative AI Applications

“Efficient similarity search, making it ideal for applications in recommendation systems, image recognition, and semantic search.”

Vector DatabaseMaker says so · source ↗
Truelink
TruelinkReal-time mobile security with AI & VPN to stop threats

“We use Qdrant to store and search high-dimensional vector embeddings of malicious patterns, phishing URLs, and threat indicators. This empowers Truelink to compare new threats against known vectors instantly using semantic AI.”

Vector DatabaseMaker says so · source ↗
Zams
ZamsThe Sales Automation Platform for B2B companies

“We evaluated a bunch of vector DBs—and Qdrant stood out for its blazing speed, filtering, and hybrid search. It's the unsung hero that lets our AI agents recall and reason across docs, CRMs, and conversations in milliseconds.”

Vector DatabaseMaker says so · source ↗
Solver
SolverOffload coding tasks to AI while you tackle bigger problems

“We love being able to self-host our custom indexes so we can be sure our customers' code is safe.”

Vector DatabaseMaker says so · source ↗
Trag
TragAI code review tool that checks every pull request against rules written in plain Englis…

“Our choice of vectordb, as engineers the experience is great!”

Vector DatabaseMaker says so · source ↗
Trieve
TrieveAll-in-one AI Infrastructure Suite

“Most performant and scaleable vector database we could find + written in Rust!”

Vector DatabaseMaker says so · source ↗
LangBot
LangBotProduction-grade platform for building agentic IM bots - 生产级多平台智能机器人开发平台/ Agent、知识库编排、插件…Vector DatabaseIn its code · source ↗

Open source: a project that declares Qdrant as a dependency in its public code — verifiable, but not necessarily a live product.

What its projects tell coding agents

36 of the 61 open-source products built with Qdrant that were worked on in the last 180 days keep instruction files for coding agents — 59%, against 53% of all the projects we read. AGENTS.md is in 25, CLAUDE.md in 20.

924words in the median file
uvthe package manager most named, in 11 of 29 files
69%say how to test a change
Named more in these files than in other projects' — in how many of the 29:
What AGENTS.md files say about other tools →What open-source products tell their coding agents →

Alternatives to Qdrant

All alternatives by situation →

Further reading

Questions makers ask about Qdrant

Is Qdrant good for production?

18 makers' products we track run on it. On GitHub, 7 public pull requests moved projects to Qdrant from tools we list, and 9 moved them away. The watch-out makers raise most: As a separate server it is slower than in-process stores for small local datasets.

What is Qdrant used for?

Open-source vector search engine written in Rust for storing and querying high-dimensional embeddings. Makers' products we track use it for vector database (18).

What are the pros and cons of Qdrant?

Pros makers name: It is fast and scales with strong price-performance, helped by its Rust core. Payload filtering and hybrid semantic plus boolean search work together. Runs easily self-hosted in Docker, a common pick for local RAG and agent memory. Cons: As a separate server it is slower than in-process stores for small local datasets. Setting up and operating a vector database is overkill for teams that just need working search.

What are the alternatives to Qdrant?

For vector database, the other options makers' products we track chose most are Pinecone (53), Weaviate (16), Chroma (12), turbopuffer (8), LanceDB (3). Open source or self-hostable among them: Weaviate, Chroma, LanceDB. Each alternative, with the situation it fits, is on the Qdrant alternatives page.

How much does Qdrant cost per month?

At 10,000 users, Qdrant costs $103 a month on the cheapest plan that fits (Standard (3 nodes, 0.5 vCPU / 4 GiB each)), from its pricing page.

Do developers switch to or away from Qdrant?

On GitHub, 7 public pull requests moved projects to Qdrant from tools we list, most from Chroma (4), Pinecone (3); 9 moved them away, most to pgvector (9).

When should you use Qdrant?

You need an open-source store with heavy filtering or hybrid search, self-hosted or managed (vector database). Something else fits better if: you already use Postgres and have up to a few million vectors, pick pgvector; you're prototyping retrieval on your laptop, or want vectors in files with no server, pick Chroma.

Is a self-hosted Qdrant secure out of the box?

No. A self-hosted instance listens on all network interfaces with no authentication until you set up API keys and TLS yourself; Qdrant Cloud clusters are secured by default. source ↗

Which languages have official clients?

Python, JavaScript/TypeScript, Rust, Go, .NET and Java. For any other language you can call the REST API or generate a client from the OpenAPI or protobuf definitions. source ↗

What are the limits of the free cloud cluster?

It is one node with 1 GB RAM, 0.5 vCPU and 4 GB disk (roughly a million 768-dimension vectors), no credit card needed. An unused free cluster is suspended after a week and deleted after four weeks of inactivity. source ↗

Can I move my data from Pinecone, Chroma or Weaviate?

Yes. The Qdrant Migration Tool, run as a Docker container, streams data in batches from other vector databases (and between Qdrant instances, such as self-hosted to Cloud) and can resume an interrupted run. source ↗

Should I create a collection per customer?

Usually not. Each collection has its own overhead, and Qdrant Cloud caps a cluster at 1000 collections by default; the recommended pattern is one collection with a tenant field in the payload, or a dedicated shard per large tenant. source ↗

Is Qdrant Cloud SOC 2 and HIPAA compliant?

Qdrant holds SOC 2 Type 2 and HIPAA certifications, with reports in its Trust Center, and offers a DPA for GDPR. Cluster data stays in the region you deploy to, while telemetry and logs go to its US management plane. source ↗

Can I skip versions when upgrading, or downgrade later?

Compatibility is only guaranteed between consecutive minor versions, so you upgrade one step at a time, and downgrades are not supported because data is migrated to the new storage format. Keep clients within one minor version of the server. source ↗

What uptime SLA does Qdrant Cloud offer?

The free tier has none. The Standard tier carries a 99.5% SLA, and Premium offers 99.9% in a single availability zone or 99.95% across several. source ↗

Is Qdrant free?

Yes — there is a free tier a small product can run on; paid use starts at Usage-based, no minimum. source ↗

Is Qdrant open source or self-hostable?

Open source, and you can self-host it. source ↗

Can AI coding agents work with Qdrant?

It serves an llms.txt docs index.

Who uses Qdrant?

18 makers' products we track, each with a source, and 94 open-source projects declare it in their code. The best known of them on Product Hunt: Lyzr (1,312 followers), Trag (1,292 followers), Superflex (1,006 followers), April (726 followers), Conva.AI (667 followers). source ↗