Couchbase Converge lets workloads built for MongoDB® or Redis® run on Couchbase with their existing drivers and clients, so you can change databases without rewriting your code. Both flavors are now in private preview, with general availability targeted for early 2027.
The inheritance problem: why teams stay on databases that no longer fit
Most teams don’t pick a database. They inherit one. Someone chose MongoDB or Redis years ago, workloads grew up around it, and now thousands of lines of code assume it’s there.
So when the database stops fitting (scaling pain, rising renewal costs, climbing compute and memory bills, or a cache that needs its own on-call rotation), the answer is usually “We’d love to move, but not if it means rewriting our code.”
And the typing isn’t the hard part anymore. An AI coding agent can crank out a rewrite in an afternoon. But then there’s the ecosystem. Your developers rarely talk to the database directly. They work through frameworks and client libraries, like Spring Data or Mongoose for MongoDB and Jedis or redis-py for Redis — plus the shells, GUIs, and IDE plugins they use every day. A traditional migration replaces all of that, too. And the new code still has to get everything else right: the edge cases, the error handling, and the business rules nobody ever documented because they live in the head of an engineer who left two years ago. Nobody wants to find out what a rewrite missed after it’s live. That’s the risk that keeps teams stuck.
Couchbase Converge lets you switch databases without the rewrite. It comes in two flavors:
- Couchbase Converge for MongoDB® Workloads runs apps built on MongoDB or Amazon DocumentDB on Couchbase.
- Couchbase Converge for Redis® Workloads runs apps built on Redis, Valkey, or Amazon ElastiCache on Couchbase.
What does Couchbase Converge change for your team?
Your code stays put
Couchbase Converge lets code built for MongoDB and Redis run on Couchbase with the drivers and clients they already use. Your developers keep the skills and tools they know. For most workloads, just the connection string changes. The code doesn’t.
You prove it before you bet on it
Converge automates the two slowest, most tedious jobs in a migration: finding what breaks and moving the data. A compatibility checker scans your code and flags anything that won’t carry over, right in VS Code or your CI pipeline. Today, it supports Java, Node.js, and Go. A data migration tool copies your MongoDB data to Couchbase and keeps it in sync, so you can run Couchbase next to MongoDB and compare the two against your real workload. Switch whenever you’re ready.
You save money sooner
With a rewrite, you keep paying for your current database, plus the engineers doing the rewrite, for as long as the rewrite takes. Converge skips the rewrite. You run both databases only while you test, and then you shut the old system down.
After the move, you run fewer systems. Couchbase handles the database and the cache on one platform, with sharding, rebalancing, and replication built in. That means fewer systems to license, patch, and staff, and more of your team’s time back for building features.
What is Couchbase Converge for MongoDB Workloads?
MongoDB is a strong document database, and plenty of teams run it happily. The trouble tends to start at scale. You have to choose a shard key (the field that decides how data splits across servers) up front, and changing it later becomes a project of its own. Scaling out costs more, too. Each new shard usually means multiple additional servers, so you add capacity in big steps instead of small ones. Rebalancing can slow down workloads while they’re running. And many teams add a separate cache because the database alone can’t hit their latency targets.
Couchbase was built for enterprise scale and speed. Data is sharded automatically. You scale by adding one node at a time. Rebalancing runs online while workloads continue serving traffic. Caching is part of the database, not a second system. And active-active cross data center replication (XDCR) lets applications write data in every region. Users work with the copy closest to them, so responses stay fast. If one region goes down, your applications keep serving users from the others.
Converge for MongoDB Workloads brings that to the code you already run, plus the frameworks and developer tools your teams use to build it.
The MongoDB ecosystem comes with you
MongoDB has one of the biggest developer ecosystems around. There are drivers for every major language and many frameworks, shells, and GUIs. These are tools your teams use every day. With Converge, those tools can talk to Couchbase. Our engineers have tested code built on Spring Boot, PyMongo, Motor, and Mongoose. Developers can keep working in mongosh and MongoDB Compass for everyday data work. Running the cluster itself, like scaling, upgrades, and security settings, happens in Couchbase’s own tools.
Your frameworks and tools stay as they are. Underneath, they pick up Couchbase’s automatic sharding, online rebalancing, built-in caching, and active-active replication. That’s what keeps the risk low. The code your business runs on stays the same. Only the engine under it changes.
If you run a MEAN stack (MongoDB, Express, Angular, and Node.js), you can keep it. As far as your code can tell, the M still looks like MongoDB: Express and Node.js still talk to it through Mongoose and the MongoDB driver. The only thing that changes is the database underneath. Your workloads never even notice.
What runs in the preview
- The MongoDB 8.0 API, with drivers for Java, Python, Node.js, Go, and .NET on wire protocol 4.0 and later
- Create, read, update, and delete, including bulk writes, cursors, and explain()
- Common aggregation pipelines for filtering, grouping, sorting, and joining data
- Single-field, compound, multikey, partial, sparse, and TTL indexes
- SCRAM and X.509 certificate authentication with built-in roles, and TLS from driver to cluster
Not in this preview yet
Multi-document transactions and change streams aren’t supported, so neither is the Kafka source connector. Secondary indexes work, but custom unique indexes don’t. Neither do time-series collections or geospatial, text, and vector search. If your code uses these, the compatibility checker flags it first. The preview isn’t certified for FIPS or FedRAMP, so regulated workloads should wait.
Sources and targets
You can migrate from MongoDB or Amazon DocumentDB 5.0 and later. Converge runs on self-managed Couchbase Server Enterprise Edition 8.0.4 or later, on Kubernetes, OpenShift, or your own hosts. The private preview runs on self-managed Couchbase only. Support for Couchbase Capella on AWS is planned for general availability.
What is Couchbase Converge for Redis Workloads?
Most people picture Redis as a fast cache in front of a database. For a lot of workloads, though, Redis is the only place the data lives. It’s the session store that keeps your users logged in. It holds the computed results and page fragments that make your site feel fast. It’s where your feature flags and configuration live. If it goes down, the applications go down with it.
Those standalone Redis workloads have a way of turning into infrastructure projects. As they grow, sharding becomes your job, with hash slots, hash tags, and hot keys to manage. Every key lives in RAM, and again in every replica. And memory isn’t getting cheaper. The same AI boom that’s squeezing GPU supply is pushing server memory prices up too. TrendForce projects server DRAM contract prices will rise about 270% in 2026.
Couchbase runs that workload differently. It grows by adding nodes, with sharding and rebalancing handled for you. You choose where data lives: fully in memory, or the hot data in memory and the rest on SSD. With the second option, you buy RAM for the data your workloads use most, not for every key you store. Either way, the data gets the same security, backup, and active-active replication as the rest of the Couchbase platform.
Couchbase Converge for Redis lets those workloads move over with the clients they already use.
Built for standalone caches first
Session stores, computed results, page fragments, feature flags, and configuration. That’s where scale and memory costs hurt most, and where Converge for Redis Workloads fits best today. A cache in front of a database works too.
What runs in the preview
- Standard Redis clients, including Jedis, Lettuce, redis-py, and ioredis
- RESP2 and RESP3 (versions of the Redis serialization protocol), pipelining, and authentication
- Strings and counters, hashes, lists, sets, and JSON
- Key expiry and scanning, so your time-to-live (TTL) settings work the way you expect
Not in this preview yet
Pub/sub and keyspace notifications are next on the roadmap. Sorted sets, Lua scripting, MULTI/EXEC transactions, blocking commands, streams, and Redis modules aren’t supported yet either.
There’s no data migration tool for Redis yet. You may not need one. When you point a cache-aside app at Couchbase, cache misses fall through to your database and repopulate the cache as traffic comes in. If Redis holds your only copy of the data, or you use it for job queues, locks, or leaderboards, talk to us first so we can check the fit.
Sources and targets
You can move workloads built on Redis, Valkey, or Amazon ElastiCache. Converge for Redis Workloads runs on self-managed Couchbase Server Enterprise Edition 8.0.4 or later, on Kubernetes, OpenShift, or your own hosts. The private preview runs on self-managed Couchbase only. Capella support will follow soon.
Can one Couchbase cluster run both your database and your cache?
Plenty of workloads run Redis in front of MongoDB. That’s two databases to license, patch, monitor, and secure for each use case.
With both flavors of Converge, the workload points its MongoDB calls and its Redis calls at the same Couchbase cluster, with no code changes on either side. Each workload gets its own bucket, with its own memory quota and access controls, so neither one crowds out the other. Your cache-aside logic, where code checks the cache first and falls back to the database, stays the same. What changes is that two platforms become one.
Who should join the private preview?
The preview is open to a limited set of customers. The best fits right now are:
- MongoDB or DocumentDB workloads that are sharded, or about to be, and do mostly create, read, update, and delete (CRUD) and aggregation work
- Standalone Redis, Valkey, or ElastiCache caches, like session stores, with scaling or memory cost pressure
- Teams that can run the compatibility checker on their MongoDB code, share a representative workload, and name an engineer to work with ours
General availability for both Couchbase Converge for MongoDB Workloads and Couchbase Converge for Redis Workloads is targeted for early 2027.
Request a spot in the private preview, or talk to your Couchbase account team about a proof of concept.
Frequently asked questions
What is Couchbase Converge?
Couchbase Converge lets workloads built for MongoDB or Redis run on Couchbase with the drivers and clients they already use. It comes in two flavors: Couchbase Converge for MongoDB Workloads and Couchbase Converge for Redis Workloads. Both are in private preview now.
Do I have to rewrite my code to use Couchbase Converge?
No. Your code keeps its drivers, queries, and commands, and you point its connection string at Couchbase. Features that aren’t in the preview yet, such as multi-document transactions in MongoDB or sorted sets in Redis, won’t carry over until they’re added.
How do I know if my MongoDB app will work?
Run the compatibility checker. It scans your code in VS Code or your CI pipeline and flags anything that won’t carry over, before you move any data. Today, it supports Java, Node.js, and Go.
Does Couchbase Converge work with Amazon DocumentDB, Valkey, or Amazon ElastiCache?
Yes. Couchbase Converge for MongoDB Workloads supports migrating from MongoDB or Amazon DocumentDB 5.0 and later. Couchbase Converge for Redis Workloads runs apps built on Redis, Valkey, or Amazon ElastiCache.
Does Couchbase Converge run on Couchbase Capella?
Not during the private preview, which runs on self-managed Couchbase Server Enterprise Edition on Kubernetes, OpenShift, or your own hosts. Support for Capella on AWS is planned for general availability of Couchbase Converge for MongoDB Workloads, with Redis Workloads to follow.
When will Couchbase Converge be generally available?
General availability for both flavors is targeted for early 2027. Join the private preview to test it on your own workload before then.
Couchbase Converge is an independent offering and is not affiliated with, sponsored by, or endorsed by the respective third-party trademark(s) mentioned. All registered trademarks remain the property of their respective owners. Use of these names is solely for the purpose of identifying product compatibility or interoperability.

Leave a comment
You must be logged in to post a comment.