Appearance
Dataplicity agent #
The on-device agent is separate from the Dataplicity CLI, which runs on an operator workstation. Neither replaces the other.
Dataplicity has two device-agent generations in the field. They share the same purpose - keeping Linux devices reachable and supportable - but they do not have the same architecture, commands or product capabilities.
For a package-level comparison, start with Choose a Dataplicity agent generation. Dataplicity OS ships a pinned modern release wheelhouse, currently 0.1.60, with rescue and supervisor services. It does not install the public legacy 0.5.x source project.
Start with your job:
| What you are doing | Use this guidance |
|---|---|
| Starting with a ready-made CM5 image | Use Dataplicity OS |
| Integrating a distribution or custom Yocto image | Integrate your Linux image |
| Building Dataplicity into a product image now | Integrate the modern agent |
| Operating or supporting a modern installation | Operate the modern agent |
| Choosing between modern and public legacy packages | Choose an agent generation |
| Reviewing what runs on a modern device | Modern agent architecture |
| Maintaining a working classic remote-access fleet | Legacy agent |
| Planning a controlled move from classic to modern | Move from legacy to modern |
A modern rescue-only installation deliberately has no supervisor. That does not make it a legacy installation. The reliable identifiers are the installed services, commands and state layout - not the hardware age.
The modern Dataplicity supervisor is a systemd service for desired-state and product functions. It is not the third-party Supervisor process manager used by some legacy installations.
Use Choose the right agent path for identification, capability boundaries and the complete documentation map.