About
Amazon Neptune is a fully managed graph database service from AWS designed for connected data workloads. It helps organizations model relationships at scale, supporting use cases such as knowledge graphs, customer 360, fraud detection, cybersecurity, and generative AI applications that benefit from graph context.
The service automatically scales graph workloads and removes the operational burden of capacity management. It also includes capabilities for graph analytics, vector search, and integrations that support GraphRAG and agentic AI workflows.
- Fully managed graph database service for connected data
- Automatic scaling for graph workloads without capacity planning
- Support for graph analytics and vector search
- GraphRAG and AI agent integrations for improved accuracy and explainability
- Enterprise security, backups, replication, and high availability features
Free Tier Value
Amazon Neptune’s pricing page says the free trial is part of AWS Free Tier and gives 750 hours of a Neptune db.t3.medium or db.t4g.medium instance for 30 days, 10 million I/O requests, 1 GB storage, and 1 GB backup storage. Using the intro_tier rule, I value the instance at the t2.micro-style benchmark ($8.50), plus EBS-like storage allowances ($3.00 for 30 GB storage and $0.00 for 1 GB backup storage is conservative here because Neptune’s own free storage is 1 GB), and I/O at the page’s $0.20 per million request example gives $2.00 for 10 million I/Os; conservatively this totals about $21.62/month equivalent. The page does not state no-credit-card-required, so that remains false.
What's included in the free tier
- 750 hours of a Neptune db.t3.medium or db.t4g.medium instance for 30 days.
- 10 million I/O requests.
- 1 GB of storage.
- 1 GB of backup storage.
See Amazon Neptune pricing for current limits.
Paid plans
Neptune Database On-Demand
- instance_hour
- per hour
- storage
- per GB-month
- I/O requests
- per million requests
- Database instances billed hourly
- Storage billed per GB-month
- I/O billed per million requests
- Supports Neptune Standard and Neptune I/O-Optimized
- Primary instances and replicas billed separately
- Multi-AZ failover supported
Neptune Serverless
- capacity
- NCU per second
- starting_capacity
- 1 NCU
- read_replicas
- up to 15
- Automatically scales capacity up or down
- Only pay when active
- Billed per second
- Starting capacity 1 NCU
- Up to 15 read replicas
- Fine-grained scaling in 1 NCU increments
Database Savings Plans
- commitment
- $/hour
- term
- 1 year
- Flexible pricing model
- Savings in exchange for commitment
- 1-year term
- Applies to eligible database usage
Pricing extracted from Amazon Neptune's pricing page. Always verify current pricing before committing.
