Buying snapshot
Scroll sideways to see all columns. With a keyboard, focus this table and use the arrow keys.
| Decision point | What Qdrant states |
|---|---|
License | Open source under the Apache License, Version 2.0 |
Self-hosted install | Run the container with |
Managed cloud | Qdrant Cloud, fully managed with high availability and auto-sharding on AWS, GCP, or Azure |
How Cloud bills | Resource usage, billed hourly |
Other products | Qdrant Edge (Beta), which runs inside your application process, and Qdrant Serverless (Waitlist) |
Which option fits
Scroll sideways to see all columns. With a keyboard, focus this table and use the arrow keys.
| Option | Price or cost basis | Choose it when |
|---|---|---|
Open source (self-hosted) | Apache-2.0 license | You will run, secure and scale the engine yourself |
Cloud Free Tier | Free forever | Testing and prototypes that fit in 1GB RAM and 4GB disk and need no high availability |
Cloud Standard Tier | Usage-based pricing | Production workloads and scaling applications |
Cloud Premium Tier | Minimum spend required | Enterprises with additional security and compliance needs |
Hybrid Cloud | Runs on your own compute, network and storage | Local data residency or regulated workloads |
Private Cloud | Dedicated, isolated deployment | Large enterprises, sensitive workloads and air-gapped setups |
What each option includes
Show full plan details
Scroll sideways to see all columns. With a keyboard, focus this table and use the arrow keys.
| Option | What Qdrant lists |
|---|---|
Free Tier | Single node cluster: 0.5 vCPU / 1GB RAM / 4 GB disk; free Cloud Inference with selected models |
Standard Tier | Dedicated resources; flexible vertical and horizontal scaling; highly available setups; backup and disaster recovery; free tokens for paid inference models; 99.5% uptime SLA; can be subscribed on a marketplace |
Premium Tier | SSO; private VPC links; 99.9% uptime SLA (99.95 with multi AZ); extra support |
Hybrid Cloud | Bring your own Kubernetes with decoupled control and data planes; data stays in your network; fully managed through Qdrant Cloud |
Private Cloud | On-premise deployment with complete control; custom SLAs; full isolation |
- Vendor source · Observed · Stored official website capture for qdrant.
- Vendor source · Observed · Stored official pricing capture for qdrant.
- Vendor source · Observed · Stored official pricing capture for qdrant.
Show full plan details
Scroll sideways to see all columns. With a keyboard, focus this table and use the arrow keys.
| Support level | Deployment option | Support price | Helpdesk and Severity 1 response |
|---|---|---|---|
Community | Open source and Free Tier | Free | Discord only; best-effort response |
Standard | Standard Tier | Included with Paid Tier | 10h/day (business hours); 4h Severity 1 response (business hours) |
Premium | Premium Tier | On Request | 24/7, 365; 1h Severity 1 response (24/7) |
- Vendor source · Observed · Stored official pricing capture for qdrant.
How Qdrant Cloud prices
Qdrant Cloud charges for the resources you use during the billing period. Usage is billed hourly, and you can watch it in the Qdrant Cloud dashboard. The bill has five parts:
- compute (vCPU) used by your clusters
- memory (GB) used by your clusters
- storage (GB) used by your clusters
- storage (GB) used by backups
- used inference tokens of paid models
A bigger cluster costs more. To size a cluster before you buy, Qdrant points to its pricing calculator.
What moves the bill
Outgrowing the free cluster. When your data grows past 1GB RAM and 4GB disk, the next step is the usage-billed Standard Tier.
Memory. Memory (GB) is a billed resource. Qdrant says asymmetric, scalar and binary quantization reduce memory usage by up to 64x while keeping search quality, so it is a lever on that part of the bill.
Backups. Backup storage is billed by the GB, apart from cluster storage.
Paid inference models. Tokens used on paid models are billed; the Standard Tier card lists free tokens for paid inference models.
Conditions to check
Hands-on engineering help is extra. A Forward Deployed Engineer is an optional add-on.
Custom-key disk encryption. Disk encryption with a custom key is available on AWS and GCP only.
Moving from self-hosted to Cloud. Qdrant provides a migration tool and documentation for OSS deployments.
Our take
If your team can run and secure its own deployment, the Apache-2.0 engine is the base option. If you want Qdrant to run it, prototype on the Free Tier, move to Standard once you need high availability, and pay Premium's minimum spend only when you need SSO or private VPC links.