Predictable pricing at any scale

Your costs are based on the resources your application uses, not how many queries you run. Scale compute, memory, storage, and GPUs as your workload grows.

Choose how you run Vespa

Every option runs the same Vespa engine. Choose based on where it runs, who operates it, and the level of control your organization needs.

  • Cloud

    Fully managed by Vespa, in Vespa Cloud

    Resource-based pricing

    vCPU, Memory, Disk, GPU memory (per hour) with volume-based discounts

    • Vespa manages infrastructure & operations
    • Automated deployment, upgrades, & scaling
    • Built-in production monitoring & security

    Start with $300 in trial credit

    Start Free Trial

  • Enclave

    Fully managed by Vespa, in your cloud 


    from $10k/mo

    Vespa fee + your cloud infrastructure spend

    • Vespa manages infrastructure & operations
    • Data stays in your cloud environment (AWS, Azure, GCP)
    • Use your cloud commitments

    Best when data must remain in your cloud environment.

    Talk to us

  • Enterprise

    Operated by your team, with Vespa support

    Custom pricing

    Annual license sized to your deployment

    • Your team manages operations
    • Runs in your cloud or data center
    • Kubernetes operator

    Best for environments where a managed service is not an option.

    Talk to us

Vespa Open Source

The full Vespa core engine, deployed and operated on your own infrastructure. Your team manages operations and upgrades, with community support available. Best for teams that want to run and manage Vespa themselves without commercial support or an SLA.

View on Github

Select the support level that fits your workload

Choose the level of support that fits your needs. Higher tiers provide faster response times, stronger operational guarantees, and more direct access to Vespa experts. You can change tiers as your needs evolve.

  • Basic

    Non-critical production workloads

    • Next business day response
    • CI validation & continuous deployment

  • Professional

    Customer-facing production workloads.

    • 1 hour response
    • 24/7 support
    • CI validation & continuous deployment
    • Backups & DR

  • Premium

    Critical workloads requiring the fastest response and named experts.

    • 15-minute response
    • 24/7 support
    • CI validation & continuous deployment
    • Backups & DR

    $20k/mo minimum

Visit the Developer Center for additional helpful resources.

Compare Deployment Options

Choose deployment based on where Vespa needs to run and how much of the operational work you want Vespa to handle.

Deployment
Open Source
Self-managed
Cloud
Fully managed
Enclave
Managed in your cloud
Enterprise
Supported self-hosted
Runs in
Your infrastructure
Vespa Cloud
Your cloud account (AWS, Azure, GCP)
Your infrastructure
Managed by
Your team
Vespa
Vespa
Your team, with Vespa support
Cloud / environment options
Cloud or on-premises
Supported Vespa Cloud regions
Your supported cloud environment
Cloud or on-premises

Core Engine
Open Source
Self-managed
Cloud
Fully managed
Enclave
Managed in your cloud
Enterprise
Supported self-hosted
Vector, lexical & hybrid search
Filtered vector search
Multi-phase ranking & reranking
Tensor computation & ML inference
Custom query processing
Grouping, aggregation, & faceting
Streaming search (personal mode)
Real-time indexing & partial updates

Platform & management
Open Source
Self-managed
Cloud
Fully managed
Enclave
Managed in your cloud
Enterprise
Supported self-hosted
Managed embedding/inference
-
-
Autoscaling controller
-
-
Application package validation in CLI
-
Performance & latency dashboards
-
Vendor-supported long-term release line
-

Operations & reliability
Open Source
Self-managed
Cloud
Fully managed
Enclave
Managed in your cloud
Enterprise
Supported self-hosted
Cluster provisioning & setup
Your team
Automated
Automated
Automated, with playbooks
Software upgrades & patches
Manual
Continuous, zero downtime
Continuous, zero downtime
Vendor supported
Multi-AZ high availability
-
Included with Basic+ Support
Included with Basic+ Support
Your team
Multi-region failover
-
Included with Premium Support
Included with Professional+ Support
Your team
Backups & disaster recovery
-
Included with Professional+ Support
Included with Professional+ Support
Your team
Monitoring & observability
-
Included
Included
Included
Uptime SLA
-
Based on support plan
Based on support plan
Based on contract

Security & compliance
Open Source
Self-managed
Cloud
Fully managed
Enclave
Managed in your cloud
Enterprise
Supported self-hosted
SOC 2 Type II
-
GDPR-aligned data residency
-
EU Regions
Your VPC
Customer-managed keys (BYOK)
-
Add-on
PrivateLink / VPC peering
-
Available
SSO / SAML
-
Audit Logging
-

Frequently asked questions

How does Vespa's $300 trial work?
You sign up without a credit card and get $300 in trial credit on a Startup-tier cluster. Use it to build, index your real data, and run real queries. When the credit runs out, the cluster tears down automatically. There is no surprise bill.
How difficult is going from trial to production?
It's the same Vespa Cloud underneath, with no migration necessary. Going from trial to a production-sized cluster is a config change in your application package. You pick a higher tier, increase node count, and redeploy.
Why not a per-query price?
Two queries can differ by 100x in the work they do. A keyword lookup against a million docs and a hybrid query that runs an ML reranker over a billion are not the same product. Pricing on resources reflects what is actually being computed.
What's the difference between Open Source and Vespa Cloud?
The core engine is the same. Cloud and Cloud Enclave add managed embedding, auto-scaling, performance dashboards, vendor-supported releases, security attestations, and a contractual SLA. Open Source is the engine; Cloud is the engine plus the operating company around it.
When does Vespa Enterprise make sense?
Vespa Enterprise makes sense when a managed service isn't an option (air-gapped, regulated, or strict on-prem environments). Vespa Enterprise is the supported way to run Vespa yourself: security-hardened image, K8s operator for automated ops, CVE and bug-fix SLAs, vendor support.