Skip to content
Matilda TrainSign up

Sovereign AI training, with your data kept in Australia.

Matilda Train checks every request against limits your administrators set, isolates each run onshore, on Australian infrastructure with no outbound internet, and records how every model was made.

Sign up Read the connection guide

Rules enforced on the server, whichever client connects

Letting an agent near GPUs is a fair worry. Matilda Train enforces the rules on the server, so they hold whichever client connects.

  • Checked before it runs

    Every request is validated against the GPU, time and resource limits your administrators set.

  • Scoped and revocable

    Grant read, submit and file access separately, and disconnect any client from your account at any time.

  • No duplicate jobs

    An agent that retries can’t launch the same run twice.

  • Isolated runs

    Each allocation has its own identity, filesystem and network isolation, with no outbound internet.

  • A record of every run

    The script digest, parameters and input checksums are captured at submission, so you can show how a model was made.

Matilda Train / Your account

Matilda Train account showing remote client setup and connected-client permissions
Connected clients and their permissions, captured from the application with an example account.
Permissions a client can ask for
research:readYour profile, run state, logs, metrics and artifacts
jobs:writeSubmit, cancel and resume your runs
files:readRead and download your workspace files
files:writeCreate files, replace or delete reviewed files, and upload datasets

Your data stays in Australia

Training data, checkpoints and run records stay on Australian infrastructure operated by Maincode. A run can’t send your data elsewhere, because allocations have no outbound internet.

When a run finishes, the weights are yours. Download them through your agent, SSH or the HTTP API, and keep them if you leave.

Where training runs
In Australia, on infrastructure that Maincode operates
Outbound internet during a run
Blocked. Each allocation has its own identity, filesystem and network isolation.
Who can see your files
Your account only. Being in the same organisation doesn’t give anyone access to your runs or files.
Where the weights end up
Your workspace, as standard files you can download and serve anywhere

Security questions

More detail, examples and troubleshooting are in the researcher documentation.

Where does my data go?
It stays on Australian infrastructure operated by Maincode. Each run is isolated, with its own identity and filesystem and no outbound internet, so training data can’t be sent elsewhere during a run. Other users, including people in your organisation, can’t see your files.
Do I own the model?
Yes. Checkpoints, adapters and merged weights are written to your workspace as standard files. Download them whenever you like and serve them anywhere. There’s no fee to keep them. The workspace guide covers downloads.
Which coding agents work?
Any client that supports remote MCP over HTTP with OAuth, including Matilda Code. You can start a client with read-only access and approve more permissions when a task needs them. The connection guide walks through it.
Can I use SSH and a desktop editor?
Yes. Start an interactive allocation, register your public SSH key, and download its SSH configuration. Connect using OpenSSH and the open-source wstunnel client on macOS, Windows or Linux. The SSH setup guide walks through it.
Do I need to know Slurm or run a cluster?
No. Your coding agent handles submission, monitoring and results through the MCP endpoint, and Maincode runs the hardware and scheduler. If you already know HPC, native Slurm is there too. See the job guide.

Bring your training.
Own the model.

Create your account, add compute credits and connect your coding agent. You pay by the GPU-hour, and every checkpoint is yours to download.

Sign up Read the getting started guide