Embodied CLI reference

An Embodied device carries the same neuraframe command as Studio, for install, licensing, health, and logs. Agent behavior (deployment mode, rules, capabilities, escalation) is set in your integration, not on the command line; see Operating & modes.

status

Service state, version, whether the engine is active or in pass-through, and license validity. On a device in pass-through, the agent serves nothing from memory and escalates, so check this first if a warm robot is escalating everything.

neuraframe status

Licensing

Per device, node-locked, with a 7-day trial. Embodied is priced separately; see Licensing & trial.

neuraframe trial start     # 7-day trial on this device
neuraframe activate        # activate a paid license
neuraframe license         # device fingerprint and validity
neuraframe license refresh # pull a renewed license
neuraframe billing         # open the billing portal

logs, doctor

neuraframe logs            # recent service logs
neuraframe doctor          # preflight plus service and license summary

uninstall

neuraframe uninstall           # keeps config, license, and learned memory
neuraframe uninstall --purge   # removes everything, including the memory pack

Without --purge the learned memory survives, so a reinstall boots warm from what the device already knew.

Memory packs are code, not a command

Exporting and importing a memory pack is done through the runtime, not a neuraframe subcommand, because it happens as part of your training and deployment flow.

MemoryPack.export(store_dir, "walker.nfmem")
MemoryPack.import_("walker.nfmem", store_dir)
Joining a fleet uses the fleet client that ships with Embodied, and is administered on the Fleet server with the separate nf-fleet command. See Embodied Fleet.