Choose a database: Cloud Firestore or Realtime Database

• Simple data is easy to store in documents, which are very similar to JSON.
• Complex, hierarchical data is easier to organize at scale, using subcollections within documents.
• Requires less denormalization and data flattening.

Learn more about the Cloud Firestore data model .

• Simple data is very easy to store.
• Complex, hierarchical data is harder to organize at scale.

Learn more about the Realtime Database data model .

• You can chain filters and combine filtering and sorting on a property in a single query.
• Queries are shallow: they only return documents in a particular collection or collection group and don't return subcollection data.
• Queries must always return whole documents.
• Queries are indexed by default: Query performance is proportional to the size of your result set, not your dataset.

• Queries can sort or filter on a property, but not both.
• Queries are deep by default: they always return the entire subtree.
• Queries can access data at any granularity, down to individual leaf-node values in the JSON tree.
• Queries don't require an index; however the performance of certain queries degrades as your dataset grows.

• Write data operations through set and update operations as well as advanced transformations such as array and numeric operators.
• Transactions can atomically read and write data from any part of the database.

• Write data through set and update operations.
• Transactions are atomic on a specific data subtree.

• A low-latency solution, with typical response times no greater than 30 ms.
• Houses your data across multiple data centers in distinct regions, ensuring global scalability and strong reliability.
• Available in regional or multi-regional configurations around the world.

• Available in regional configurations. Databases are limited to zonal availability within a region.
• Extremely low latency, with typical response times no greater than 10 ms. An ideal option for frequent state-syncing.

• Typical uptime performance of 99.999%.
• If availability is of the utmost importance, for example in ecommerce apps, use Cloud Firestore .

• Typical uptime performance of 99.95%.

• Scales completely automatically. Scaling limits are around 1 million concurrent connections and 10,000 writes/second. We plan to increase these limits in the future.
• Has limits on write rates to individual documents or indexes.

• Scale to around 200,000 concurrent connections and 1,000 writes/second in a single database. Scaling beyond that requires sharding your data across multiple databases.
• No local limits on write rates to individual pieces of data.

• Reads and writes from mobile SDKs secured by Cloud Firestore Security Rules .
• Reads and writes from server SDKs secured by Identity and Access Management (IAM) .
• Rules don't cascade unless you use a wildcard.
• Rules can constrain queries: If a query's results might contain data the user doesn't have access to, the entire query fails.

• Reads and writes from mobile SDKs secured by Realtime Database Security Rules .
• Read and write rules cascade.
• You validate data separately using the validate rule.

Cloud Firestore supports daily spending limits for App Engine projects, to make sure you don't go over the costs you're comfortable with.

Read more about Cloud Firestore pricing plans .

Read more about Realtime Database pricing plans .