Store
Give your data a home.
Upload objects to your namespace. Immutable versions preserve the object behind each write, while paths give you a familiar way to organize it.
m| object storeClosed alpha
Move the compute
to the data.
Object storage with compute beside it. Store your data once, then bring parallel jobs to the objects you already have.
Storage meets compute
Keep the data close.
Bring your own tools.
m carries forward the idea behind Manta: run programs where the data lives. Object storage and parallel processing share a workflow, from the first upload to the final result.
Compute
Take the work to the data.
Run parallel jobs beside stored objects. Use familiar Unix tools and connect map and reduce stages to process the data where it lives.
Use
Bring back the result.
Inspect job outputs, retrieve the results, or use them in the next stage of your workflow. Keep working from your terminal, application, or agent.
Choose your way in
One set of objects.
Familiar ways to work.
Use the terminal, an API, a filesystem mount, or an agent. Each interface connects to the same stored objects, with its own documented capabilities.
Explore the technical overview- CLI
- Upload objects, explore paths, and submit jobs from your terminal.
Command line - REST
- Work with objects and jobs through the HTTP API.
API reference - S3
- Connect S3 clients using the supported compatibility profile.
Compatibility profile - Filesystem
- Browse your namespace through a local NFS gateway.
Filesystem access - MCP
- Give connected agents an interface to storage and compute.
MCP for agents
Help shape what
storage can do.
m is in closed alpha. Tell us about your data, your workload, and the tools you use. We can discuss whether the alpha is a fit.
Discuss alpha access Alpha access is limited. This is an invitation to start a conversation.
Built at L Labs