SHIP APPS AND AI AGENTS FASTER, NO DATABASE TO RUN
Couchbase Capella
Capella is Couchbase, fully managed on AWS, Google Cloud, and Azure. Your team ships features instead of patching clusters, and your app stays fast when traffic peaks.
Your database shouldn’t need its own team
Running a distributed database eats engineering time, and stitched-together stacks break at peak. Capella takes both problems off your plate.
Protect revenue on your busiest day
Slow pages lose carts. A memory-first engine keeps reads fast at peak, backed by a 99.99% uptime SLA on production plans.
Get AI features into production
Agents need memory, tools, and fresh data. Capella on AWS runs all three next to your live data, so pilots don’t stall on plumbing.
Put engineers back on the roadmap
Couchbase handles patching, upgrades, backups, and 24/7 monitoring. Your engineers stop babysitting clusters and ship features.
Know what you’ll pay
Pay per node per hour, scale each service independently, and switch idle clusters off. Billing alerts flag surprises early.
What is Couchbase Capella?
Capella is the fully managed cloud service for Couchbase. It’s the same engine as Couchbase Server, run for you on AWS, Google Cloud, and Azure.
Capella runs Couchbase Server, the same engine teams run on premises: JSON, key-value, SQL, full-text and vector search, eventing, and a built-in cache on one copy of data. Your code, SDKs, and skills carry over both ways, so choosing managed today doesn’t lock you out of running it yourself tomorrow.
Couchbase deploys, patches, upgrades, backs up, and monitors your clusters, and Health Advisor flags trouble before it pages anyone. You pick the cloud, region, node size, and services, and you keep control of data, users, and access. Everything in the UI is also in the API and Terraform.
On Capella on AWS, the AI Data Plane adds managed Agent Memory, an MCP Server, a Model Service with semantic caching, AI Functions in SQL, and automatic vectorization. Agents get memory, tools, and models next to your operational data. You don’t wire up a separate vector database, cache, or memory store.
Capella Analytics runs real-time analytics on fresh JSON, with no ETL pipeline to babysit. Point it at Capella, Kafka, or CDC from Oracle, then query with SQL++ or your BI tool. Capella App Services syncs data to mobile and edge apps that keep working offline.
Built to stay up, wherever your users are
99.99%
uptime SLA on Developer Pro and Enterprise clusters
85+
cloud regions across AWS, Google Cloud, and Azure
CUSTOMER RESULTS
Teams that stopped running databases
UNDER THE HOOD
Everything a production database needs, run for you
Everything Capella handles for you, plus new agent services
- Operate
- Scale
- Secure
- Automate AI
- Connect
Couchbase patches, upgrades, and watches your clusters. You get health checks, alerts, and an API for everything.
Key capabilities, including what’s new on Capella
Managed patching and upgrades
Couchbase patches and upgrades clusters in windows you schedule. Each Server release gets 36 months of patches and 6 months’ notice before retirement.
Backups that restore in minutes
Schedule full-cluster snapshots that restore with indexes, typically within minutes, or back up single buckets with full and incremental runs.
Terraform, APIs, and Capella iQ
Everything in the UI is also in the REST API and Terraform. Capella iQ turns plain English into SQL and suggests indexes.
Scale each service on its own, let storage grow by itself, and turn clusters off when nobody’s using them.
Key capabilities, including what’s new on Capella
Big data, small memory
The Magma storage engine holds up to 16 TB per node and can serve 5 TB from 64 GB of RAM, so big datasets don’t force huge nodes.
Capella Express Scaling
New in private preview on AWS: add capacity for traffic spikes and disaster recovery in minutes instead of hours. Ask your account team to join.
Turn clusters off
Pause dev and test clusters on a schedule or by API for up to 30 days. Data, indexes, and settings stay put while you pay a lower off rate.
Encryption you can’t turn off, your own keys if you want them, private networking, and audited compliance.
Key capabilities, including what’s new on Capella
Private networking
Connect over AWS PrivateLink, Azure Private Link, Google Cloud Private Service Connect, or VPC peering, and keep traffic off the internet.
Encryption that’s always on
TLS 1.2 or higher on every connection, and it can’t be turned off. Data at rest uses AES-256, with keys held in FIPS 140-2 validated HSMs.
Audited compliance
SOC 2 Type II, ISO 27001, PCI DSS, HIPAA, and CSA STAR Level 2 cover Capella, so the paperwork is done before you ask.
New: managed Agent Memory and MCP Server. Plus models, AI Functions, and vector search next to live data.
Key capabilities, including what’s new on Capella
Managed Agent Memory
Agents remember users and past sessions, whether you build with LangGraph, CrewAI, LlamaIndex, or Strands. Fully managed on AWS.
Managed MCP server
Agents query and act on Capella data through MCP. Couchbase scales it inside limits you set. Fully managed on AWS.
Private vectorization
Vectorize data with your own Amazon Bedrock models over a private connection, so content never crosses the public internet. On AWS.
Replicate across regions and out to self-managed clusters, reach data over REST, and analyze it without ETL.
Key capabilities, including what’s new on Capella
Hybrid with Couchbase Server
XDCR (cross data center replication) connects Capella and self-managed Couchbase Server for migrations, hybrid apps, or a copy in your own data center.
Capella Analytics
Stream in data from Capella, Kafka, or CDC on Oracle and SQL Server, then query with SQL, Tableau, or Power BI. No ETL jobs.
Capella App Services
Sync data to Couchbase Lite apps on phones and edge devices. Apps keep working offline and catch up when they reconnect.
- Global
- AI agents
- Analytics
ARCHITECTURE
Serve users from the closest region
Users expect your app to work wherever they are. Capella runs active clusters in regions near them, kept in step by XDCR in both directions. If a region goes down, traffic moves to one that’s already live. Nexon launches a new game region in 20 minutes this way.
ARCHITECTURE
Agents that remember and act
Most agent stacks bolt a vector database, a cache, a memory store, and a model gateway onto the app database. That’s five systems to secure. On Capella, they share one platform with your live data, so you tighten your security perimeter and agents answer from what’s true now.
ARCHITECTURE
Live analytics without ETL jobs
Analytics on a nightly copy answers yesterday’s question. Capella Analytics streams in data from Capella, Kafka, and CDC sources like Oracle, stores it in columns and runs SQL in parallel. Results flow to Tableau or Power BI, or straight back into your app.
FREE TIER
Free forever, no credit card
The Capella free tier gives you a real operational cluster, not a sandbox: one node with 8 GB, SQL, key-value, vector and full-text search, and App Services for mobile sync. You don’t need a credit card, and it stays free as long as you keep using it. When you’re ready for production, deploy a paid cluster and keep the same code.
Same engine, three ways to run it
MANAGED CLOUD
Capella, in four plans
Free is for learning. Basic suits development and apps that aren’t mission-critical. Developer Pro adds Capella Support within 8 hours and a 99.99% uptime SLA. Enterprise adds 24/7 support with a 30-minute response, plus your own keys and audit logs.
SELF-MANAGED
Couchbase Server
Run the same engine yourself, on premises, in any cloud or on Kubernetes with the Autonomous Operator. It’s the right fit when you need full control of the infrastructure, or when your data can’t leave your data center.
MOBILE AND EDGE
Couchbase Mobile
Couchbase Lite puts the database on phones, kiosks, and edge devices, so apps keep working without a connection. Capella App Services keeps those devices in sync with the cloud and with each other.
Couchbase Capella FAQ
Answers to common questions about Capella, its pricing, and how it compares to Couchbase Server.
Capella is a fully managed cloud database service from Couchbase. It stores JSON and serves key-value, SQL, search and vector workloads, on AWS, Google Cloud, and Azure.
You pay by the hour for each cluster, based on cloud, region, node count, CPU, RAM, and storage. Plans are Free, Basic, Developer Pro, and Enterprise. Pay by card, prepaid credits, or as you go.
Capella runs Couchbase Server, today versions 8.0 and 7.6. You operate Server yourself. Couchbase operates Capella for you, handling setup, scaling, backups, and upgrades.
Yes. Each organization gets one free tier operational cluster with 8 GB, SQL++, vector search, and App Services. It needs no credit card and stays free while you use it.
Yes. Capella runs vector search on live data. On AWS, it adds managed Agent Memory and an MCP Server, both new, plus a Model Service, AI Functions, and vectorization.
Yes. TLS and AES-256 encryption are always on, Enterprise adds your own keys, and Capella is audited for SOC 2 Type II, ISO 27001, PCI DSS, HIPAA, and CSA STAR Level 2.
Keep exploring
-
Capella in two minutes
A quick tour of the console, from first cluster to first query.
-
Capella security and data protection
Encryption, access control, compliance, and who is responsible for what.
-
Intrigued? Start today.
Let’s talk about how you can build smarter for an AI future, today.