Skip to main content
Infrastructure is the foundation of Meilisearch Cloud. Every search experience you build starts with a project: a fully isolated Meilisearch instance with its own indexes, API keys, settings, and resource allocation.

Projects

A project is the core unit of Meilisearch Cloud. Think of it as a dedicated Meilisearch server:
  • Each project has its own indexes and documents
  • Each project has its own API keys and security settings
  • Each project is either resource-based or usage-based (see below)
You can create multiple projects in the same account, for example one per application or one per environment (production, staging).

Resource-based vs usage-based projects

When creating a project, you choose a billing model: Resource-based projects give you a fixed allocation of CPU, RAM, and storage. You have full control over the resources available to your instance, and billing is predictable: you pay for what you provision, regardless of traffic. Usage-based projects scale automatically. Instead of choosing a resource tier, you pay based on the number of searches performed and documents indexed. Meilisearch Cloud provisions the resources needed to serve your workload without you managing capacity. This model is simpler to get started with, but costs are harder to predict since they vary with your usage.
Resource-basedUsage-based
BillingFixed resources (CPU, RAM, storage)Searches and documents
ScalingManual (you choose the tier)Automatic
Price predictabilityHighDepends on traffic
Best forProduction workloads with known trafficVariable or early-stage workloads

Regions

Meilisearch Cloud is available in multiple regions across the world. Choose the region closest to your users for the lowest latency.
CodeLocationRecommended for
FRAFrankfurt, GermanyEuropean users
LONLondon, United KingdomUK and Western Europe
SGPSingaporeSoutheast Asia and Oceania
JPNJapanJapan and Northeast Asia
SFOSan Francisco, United StatesUS West Coast
NYCNew York, United StatesUS East Coast and global default

Resource tiers

Each project runs on a resource tier that defines its CPU, RAM, and storage allocation. Choose a tier based on:
  • Index size: Larger indexes require more RAM for fast search
  • Query volume: High QPS benefits from more CPU
  • Vector search: Storing and searching vectors requires additional RAM proportional to the number of dimensions and documents
You can change the resource tier at any time from the project settings.

Premium Compute

Premium Compute is a class of dedicated infrastructure available on Enterprise plans. Compared to standard resource tiers, Premium Compute offers:
  • Dedicated physical resources (no shared CPU or memory with other tenants)
  • Higher-performance CPUs with better single-thread and multi-thread throughput
  • Faster disk I/O for index reads and writes
  • Optimized builds tuned for search workloads
Premium Compute is suited for production workloads with strict latency requirements or very large indexes. Contact sales@meilisearch.com to enable it.

Next steps

Create a project

Step-by-step guide to creating your first project

Regions

Detailed region reference and latency guidance

Manage resources

Scale CPU and RAM to match your workload

Version management

View and upgrade your Meilisearch version

Sharding and replication

Scale to large datasets with high availability

Backups

Automatic weekly backups with Enterprise customization options