Typed artifacts
Containers, ONNX and TensorRT model payloads, placed configs and systemd service binaries have distinct activation and recovery rules.
Edge AI delivery
OTA updates for Linux edge AI fleets
Meshanics treats models and the software around them as one release problem. Artifacts are digest-pinned, published through TUF, targeted by device capability, and rolled out through observable waves.
Good fit
When to evaluate it
- +Jetson, Raspberry Pi and x86 devices in the same product fleet
- +Teams shipping containers and model weights on different cadences
- +Products that need customer-owned, on-premise or air-gapped delivery
Product boundary
What the current product does
Capability targeting
Release constraints can match architecture and hardware profile instead of assuming one golden image for every edge device.
Quality evidence
The ML SDK can submit bounded aggregate windows so operators compare model-version quality without uploading raw samples.
One release record
Signed release manifests fix the ordered artifact set, and device-signed receipts record the result of customer-authorized releases.
Explicit limit
Meshanics is not a model registry, experiment tracker, or full host-monitoring product. It delivers approved artifacts and records bounded rollout and model-quality facts.
Console proof
See the operational record
Questions
Does Meshanics require changes to application code?
No. The static Linux agent installs on the device and activates signed payloads through typed adapters. Health behavior and recovery limits are configured for the artifact, not embedded in customer application code.
Does Meshanics certify CRA compliance?
No. Meshanics records and exports evidence for the manufacturer's compliance work. The manufacturer remains responsible for product scope, risk decisions, conformity assessment, reporting, and declarations.