cpuos

Base profile

Custom image sandbox template

Design a custom sandbox image with a pinned toolchain, reproducible build inputs and a smoke test before turning it into a reusable template.

When to use this profile

Choose a custom image when preparation dominates execution or a task needs software outside the base profiles. Build the toolchain once, keep user data and credentials out of the image, and run a deterministic smoke test before publishing a new version. An OCI build is a way to describe the filesystem; it does not by itself implement a microVM execution service.

Prepare a repeatable environment

  • Pin the base image by digest and record package versions.
  • Keep the image free of tokens, private datasets, SSH keys and machine-specific configuration.
  • Define a smoke test that checks binaries, filesystem permissions and the expected command.
Command inside a prepared Linux guest
sh /work/smoke-test.sh

Return results the agent can use

  • An image digest and build manifest.
  • The smoke-test result for that exact version.
  • A documented input, output and network contract.

Rebuild from a clean environment, compare the recorded inputs, and execute the smoke test without host mounts or production credentials.

Resources and boundaries

Start with 2 vCPU and 4 GB of RAM, then measure peak memory and task duration on a representative fixture. These are workload planning values, not a benchmark or a provisioned configuration. Use the sandbox cost calculator to estimate running time and retained snapshots.

  • Custom image support depends on the runtime adapter and image conversion pipeline.
  • A large image increases storage and preparation costs; retain only dependencies needed at runtime.
  • Rebuild and retest after security updates rather than treating an old snapshot as permanently safe.

Keep model inference separate from this execution profile. A hosted model or gpuOS can decide the next action while the CPU environment runs it. The quickstart describes the account workflow and the proposed runtime contract.

Questions

Is this a separate built-in template?
This page documents the custom base profile shown in the cpuOS product catalog. Runtime provisioning requires a connected execution service; creating an account does not provision a microVM.
Can I run the command now?
The command runs in a Linux environment where the listed dependencies and input files are prepared. cpuOS SDK examples describe a proposed contract; confirm runtime access and package versions before integration.
How should I choose CPU and memory?
Use the starting profile to run a representative fixture, measure peak memory and elapsed time, and add room for package installation, worker processes and larger inputs. Enforce a task timeout separately.

Related guides

Plan your custom image task

Create a workspace and choose a profile. Connect an execution backend before running code.

gpuOS · where models think

Need the model too? Run it on gpuOS

gpuOS serves open models on your own GPUs behind one OpenAI-compatible API. The model reasons on gpuOS, the agent acts in a cpuOS sandbox.