Inventory & telemetry
Inspect connectivity, reported software, telemetry and device events, grouped by organisation and model.
Open-source device management
Anchor. Keep your fleet within reach.
Bring device inventory, firmware releases and remote operations together in a server you can host yourself.
Built by Julien Vermillard at ClunkyMachines, from the recurring engineering work of connecting, maintaining and updating products in the field.
Technical preview · AGPL-3.0 · Read the v0.1.0 launch story
The problem
After the first connection, the real work starts
Shipping a device is the beginning of its operational life. Teams still need to answer which software is running, why a device stopped reporting and whether a corrective update reached the right units.
Anchor brings that repeated infrastructure into one project, so engineers can spend more time on the product itself.
Capabilities
Devices, releases and operations in one place
Tasks & campaigns
Run read, write or firmware-update tasks for a device or selected fleet, then track results.
Releases & vulnerabilities
Associate firmware with SPDX SBOMs. Grype supplies CVE scanning; engineers review relevance to the product.
Capability reference: project README , checked on 9 October 2026. Screenshots show the v0.1.0 launch context, not every subsequent development.
Architecture
A clear boundary between the fleet and the device
Connected devices
MQTT through Mosquitto, or CoAP over DTLS with a separate frontend. Devices report state and execute tasks.
Anchor server
A Go application with a web interface, REST API and SQLite by default. Releases and operations are managed centrally.
Your engineering system
Connect manufacturing and release pipelines to fleet operations. Keep signing authority, deployment approval and support ownership explicit.
The device remains responsible for a safe update
When reviewing an integration, we trace the whole path: build, sign, publish, download, validate, boot, confirm and report. A successful server-side task alone cannot prove image authenticity or recovery after an interrupted boot.
For a Zephyr product, this means examining the actual MCUboot configuration and health checks alongside the backend. For other stacks, the same questions apply to the mechanisms they provide.
Product security
Connect a release finding to the fleet
A useful vulnerability response starts with traceability: the component in a build, the release containing it and the devices reporting that release.
Your team then decides whether the finding affects the product, prepares a correction and checks the outcome in the field. Preserve the assessment and deployment evidence alongside the release.
Getting started
Try it with a disposable fleet first
1. Choose your build
Read the deployment instructions and select a tagged release or a recorded source revision. Keep that version with your trial notes.
View releases2. Connect a device
Choose MQTT or CoAP, follow the matching protocol contract and confirm heartbeat, version reporting and task results.
Connection documentation3. Exercise the workflow
Use the documented fleet simulator for a local trial. Test an update, a failed task and an offline device before integrating hardware.
Fleet simulator guide
Maturity & limits
Technical preview, with work still ahead
Use a trial to establish what your product needs from fleet operations. Integration, deployment hardening and validation must be assessed for your environment.
Make Anchor work with your product
I can help connect firmware, provisioning, release tooling and infrastructure, then validate the operational workflow with your team.