Deploy and Host Qdrant

Deploy a self-hosted Qdrant vector database on Railway in one click. This template provisions the official qdrant/qdrant Docker image with persistent storage pre-configured at /qdrant/storage, so your vector data survives restarts without any manual setup.

For quickstart refer following guide - Guide

To access the dashboard, go to your-railway-url.app/dashboard:

About Hosting Qdrant

Qdrant is an open-source, high-performance vector database and similarity search engine written in Rust. It's purpose-built for storing, indexing, and querying high-dimensional vector embeddings — the kind generated by OpenAI, Cohere, or sentence-transformers.

Key features:

• HNSW indexing for sub-millisecond ANN (approximate nearest neighbour) search
• Rich payload filtering — combine vector search with structured metadata conditions
• Scalar, Product, and Binary Quantization — reduce memory usage up to 40x
• REST and gRPC APIs with OpenAPI v3 spec; official clients for Python, TypeScript, Rust, Go
• Apache 2.0 licensed — no vendor lock-in

Why Deploy Qdrant on Railway

Railway handles the infrastructure so you can focus on building. Compared to managing Qdrant on a raw VPS, Railway gives you automatic TLS, environment variable management, volume persistence, and one-click redeployments — all from a clean UI. Private networking means your app services can reach Qdrant internally without exposing it publicly. The free tier is enough to prototype, and scaling up is a slider, not a support ticket.

Common Use Cases

• Retrieval-Augmented Generation (RAG): Store document embeddings and retrieve context for LLM prompts with LangChain or LlamaIndex
• Semantic search: Power search that understands meaning, not just keywords — across products, articles, or support docs
• Recommendation systems: Use Qdrant's Recommendation API to surface personalised suggestions based on vector similarity
• AI agents: Give agents long-term memory by persisting and querying embeddings across sessions
• Anomaly detection: Compare incoming data vectors against known-good baselines in security or monitoring pipelines

Dependencies for Qdrant

Qdrant is self-contained — no external database or broker required.

• Your application service (any language) connecting to Qdrant via REST or gRPC

Environment Variables Reference

Deployment Dependencies

• Docker image: qdrant/qdrant (GitHub: qdrant/qdrant)
• Persistent volume mounted at /qdrant/storage — already configured in this template
• No other services required

Getting Started with Qdrant After Deployment

Once your Railway deployment is live, the REST API is available at your Railway-assigned public URL on port 6333. Test it immediately:

curl https://your-app.railway.app/collections \
  -H "api-key: YOUR_API_KEY"
# Returns: {"result":{"collections":[]},"status":"ok"}

Create your first collection and insert a vector using the Python client:

from qdrant_client import QdrantClient
from qdrant_client.models import Distance, VectorParams, PointStruct

client = QdrantClient(
    url="https://your-app.railway.app",
    api_key="YOUR_API_KEY"
)

client.create_collection(
    collection_name="docs",
    vectors_config=VectorParams(size=1536, distance=Distance.COSINE)
)

client.upsert(
    collection_name="docs",
    points=[PointStruct(id=1, vector=[0.1] * 1536, payload={"text": "hello"})]
)

Install with pip install qdrant-client. From here, connect your embedding pipeline and start querying.

Hardware Requirements for Qdrant

Qdrant's memory usage scales with your vector dimensions and collection size. As a baseline:

• Development / prototyping: 1 vCPU, 512MB RAM — fine for collections under 100k vectors
• Small production workloads: 2 vCPU, 2–4GB RAM — handles millions of 768–1536-dimension vectors with quantization enabled
• High-throughput production: 4+ vCPU, 8GB+ RAM — for unquantized large collections or heavy concurrent query loads

Enable Binary or Scalar Quantization to cut RAM requirements by 4–40x on large datasets.

Qdrant vs Competitors

Qdrant vs Pinecone: Pinecone is fully managed and easier to start with, but costs scale fast and you have no control over your data. Self-hosted Qdrant on Railway gives equivalent search quality at a fraction of the cost.

Qdrant vs pgvector: If you're already on Postgres and have under 1M vectors with simple filtering needs, pgvector works. For dedicated vector workloads, larger collections, or quantization, Qdrant is significantly faster.

Self-Hosting Qdrant (Outside Railway)

To run Qdrant on your own machine or VPS using Docker:

docker pull qdrant/qdrant

docker run -d -p 6333:6333 \
  -e QDRANT__SERVICE__API_KEY=your-secret-key \
  -v $(pwd)/qdrant_storage:/qdrant/storage \
  qdrant/qdrant

Access the REST API at http://localhost:6333. Data persists in ./qdrant_storage. For production VPS deployments, wrap this in Docker Compose and add a reverse proxy with TLS.

Is Qdrant Free to Use?

Qdrant is fully open source under the Apache 2.0 license — free to self-host with no usage limits or feature gates. Qdrant Cloud (the managed offering) has a free tier with limited capacity; paid plans start around $25/month. On Railway, you pay only for the infrastructure you use — typically a few dollars per month for small workloads.

FAQ

Is the setup process complicated? No. This Railway template requires minimal configuration — deploy, optionally set an API key, and your Qdrant instance is live. The volume is pre-mounted so data persistence works out of the box.

How do I run Qdrant with Docker locally? Pull the image with docker pull qdrant/qdrant, then run it with port mapping and a volume for persistence: docker run -p 6333:6333 -v $(pwd)/qdrant_storage:/qdrant/storage qdrant/qdrant. The API is then available at localhost:6333.

What is Qdrant used for? Qdrant powers advanced semantic search, recommendation systems, retrieval-augmented generation (RAG), data analysis, anomaly detection, and AI agent memory — any use case that involves finding similar vectors in high-dimensional space.

Should I use Qdrant Cloud or self-host? Qdrant Cloud is convenient but expensive at scale. Self-hosting on Railway gives you full control over your data, no vendor lock-in, and predictable infrastructure costs — with far less operational overhead than managing your own VPS.

How do I secure my Qdrant instance? Set the QDRANT__SERVICE__API_KEY environment variable before your deployment goes public. All client requests must then include an api-key header. On Railway, you can also restrict access to internal private networking so only your app services can reach Qdrant directly.