# Fleetwide > Fleetwide delivers over-the-air updates to Docker containers that run where you can’t reach them directly: customer sites, on-prem servers, edge devices and AI agents. Whoever runs the host starts the Fleetwide supervisor image, one container, the same way they would start yours, and it stays on the release you publish; nothing is installed on the host or set up in the cluster. Fleetwide can also be built into your own image. Publish a release once; every container pulls it, checks its own health and rolls back on its own, and the console shows every fleet, its version and its health. Fleetwide is a hosted console plus a small supervisor. A vendor publishes a digest-pinned release of a Docker image and promotes it to a channel; every container that follows the channel updates itself, reports health and metrics, and rolls back if the release is unhealthy. Over-the-air updates and fleet management for Docker containers. ## Use cases - AI agents you ship to customers: Prompts, tools and model choices change every week. Publish a release, roll it to a few containers first, and any container that fails its health check or an error-rate gate rolls itself back. - Software installed on-prem: Your product runs in customer data centers and VPCs you cannot sign in to. Each container pulls releases from your registry, public or private, and updates only inside that customer’s window, in their time zone. - Monitoring & security agents: Keep collectors and sensors on every host on the current version. Roll a fix out a batch at a time, watch the versions converge, and see exactly which hosts are behind. - Edge & IoT devices: Gateways, kiosks, Jetsons and Pis running Docker or k3s. Only changed layers download, the current release keeps running while a device is offline, and updates can wait for a quiet hour. - Models, prompts & config: Ship model weights, prompt files and rules as versioned assets from any OCI registry or Git repository, without rebuilding the image. Each one swaps atomically, then restarts the app, signals it or runs a command. - Locked-down clusters: Where only non-root containers may run, such as OpenShift, the supervisor is built into your image. No second container, no root, no added capabilities, and the console says when a release needs a full redeploy. - Development & test fleets: Point a fleet at your test machines with logs and proxy switched on. Push a build, stream its logs live, and open the running container in your browser through Fleetwide Proxy, with no ports opened. ## Plans - Dev (10 containers, $0 / forever): 7-day metrics retention; Unlimited apps, releases and images; Analytics and telemetry; Public Fleetwide Proxy on fleetwide.app - Team (100 containers, Early access*): 30-day metrics retention; Fleetwide Proxy behind your organization’s sign-in; Custom domains for Fleetwide Proxy; Everything in Dev; 30-day free trial, no credit card - Enterprise (Unlimited containers, Talk to us): Unlimited metrics retention; Self-hosting; Fleetwide Bridge: secure private tunnels; Everything in Team - Pricing: https://fleetwide.io/#pricing ## Docs - [How Fleetwide works](https://fleetwide.io/docs): You publish a release. Every container that follows the channel pulls it, checks its own health and rolls back on its own. - [Quickstart](https://fleetwide.io/docs/quickstart): The fastest way to see it work: on the Welcome page, Try it with a sample app creates a sample app, release, fleet and deployment in one click and shows the… - [Apps and releases](https://fleetwide.io/docs/releases): An app is one image you ship: its key, registry and repository, the assets it carries, its runtime (supervisor or embedded), its health check and the metrics… - [Channels and promotion](https://fleetwide.io/docs/channels): Fleets follow a channel: stable, beta, whatever you name. Promoting a release to a channel is what makes containers move; they pick it up on their next… - [Registries and credentials](https://fleetwide.io/docs/registries): The registry belongs to the app. Public: the console resolves the tag to a digest itself and checks that the digest you give, if any, matches. - [fleetwide.yaml: health and readiness](https://fleetwide.io/docs/manifest): Ship a /fleetwide/fleetwide.yaml in your image. The health check decides whether a release is good, so a release that cannot pass it never becomes the one… - [Fleets, deployments and containers](https://fleetwide.io/docs/fleets): A fleet is a customer, a site or an environment: it follows one channel and carries its own variables, credentials, update window, rollout settings and… - [Install on a customer](https://fleetwide.io/docs/install): From a fleet, Add deployment names the deployment (labels and the ephemeral setting under Advanced), mints its first enrollment key and shows the run command… - [Kubernetes and k3s](https://fleetwide.io/docs/kubernetes): One Kubernetes Deployment is one Fleetwide deployment. The pod runs the supervisor image with the enrollment key from a Secret and is known by its pod. - [Rolling updates](https://fleetwide.io/docs/rollout): The deployment is the unit of rollout. When the fleet's channel moves, only so many of a deployment's containers are updated at once: the fleet's setting, or… - [Rollback and the failure policy](https://fleetwide.io/docs/rollback): A new release is pending until it has been healthy for grace + 2×interval. - [Enhanced security mode](https://fleetwide.io/docs/embedded): Clusters that forbid root (Pod Security restricted, OpenShift) cannot run the supervisor image, which writes your image over the container's filesystem. - [Install the CLI](https://fleetwide.io/docs/cli-install): The fleetwide CLI builds embedded images and prints digests. It is one binary with no dependencies, published on GitHub for macOS and Linux, amd64 and arm64. - [Tag and push](https://fleetwide.io/docs/cli-tagging): --tag is where the finished image goes. It is required, and you give at least one of --push or --output. - [fleetwide embed](https://fleetwide.io/docs/cli-embed): Builds the supervisor into your image. With --api-key and --app it reads the app's runtime, sync paths and assets from the console and makes those paths… - [fleetwide digest](https://fleetwide.io/docs/cli-digest): Prints what a release should pin for an image the console cannot reach. - [Monitoring: containers, resources and health](https://fleetwide.io/docs/monitoring): Monitoring, under Run, is every container of the organization in one place, in four tabs. - [Telemetry and metrics](https://fleetwide.io/docs/telemetry): Every check-in carries the container's state: phase, health, restarts, the last exit (code, signal, out of memory), and CPU, memory and state-volume usage… - [Live logs](https://fleetwide.io/docs/logs): With the fleet's live logs switch on and the developer supervisor image, open a container or a deployment and switch to Logs: the app's output streams to the… - [Fleetwide Proxy](https://fleetwide.io/docs/proxy): A proxy opens one container port of a deployment at https://.fleetwide.app. - [Custom domains](https://fleetwide.io/docs/domains): Custom domains (Proxy → Custom domains, Team plan) serve a proxy at your own hostname. - [Fleetwide Bridge](https://fleetwide.io/docs/bridge): Fleetwide Proxy lets you reach into a container. Fleetwide Bridge is the other direction: it lets the container reach out, to something on the customer's own… - [Models, assets and variables](https://fleetwide.io/docs/assets): Assets (Publish → Assets) are versioned files that ship beside the image without a rebuild: model weights, prompt files, rule sets. - [Edge and IoT](https://fleetwide.io/docs/edge): Fleetwide updates the containers on a device, not its operating system. Anything running Docker or k3s works: gateways, kiosks, Raspberry Pi, Jetson. - [Plans](https://fleetwide.io/docs/plans): Plans are priced by containers: one enrolled supervisor is one container, whatever deployment it belongs to; a Kubernetes pod is one container.