Bolt is an open source, agentless orchestration tool that automates the manual work of maintaining infrastructure on demand or as part of a larger orchestration workflow, and it is built for infrastructure and operations engineers who manage Linux, Windows, and macOS targets.

What it is

Bolt is an open source orchestration tool maintained under the puppetlabs organisation and written in Ruby, released under the Apache-2.0 licence. It sits in the Infrastructure & Operations / Orchestration & Scheduling category and is tagged with bolt, devops, orchestration, puppet, ssh, and winrm. It is installed on a local workstation and connects directly to remote targets over SSH or WinRM, so no agent software is required on the nodes being managed. The documentation lives on the Bolt website, and support runs through the #bolt channel on the Puppet community Slack.

The problem it solves is the manual work that infrastructure maintenance demands. Rather than logging into servers one at a time to patch and update systems, troubleshoot servers, deploy applications, or stop and restart services, an operator expresses that work once and lets Bolt run it. The same tool covers tasks performed on an as-needed basis and tasks that form part of a greater orchestration workflow, and plans can be created and shared to expand that automation across an application stack. Existing scripts and plans in YAML, PowerShell, Bash, Python, or Ruby can be brought along, and content from the Puppet Forge can be reused.

Key capabilities

• Agentless remote execution: installs on a Linux, Windows, or macOS workstation and connects to targets with SSH or WinRM, with no agent software on the managed nodes.
• Bolt plans for orchestration: run simple plans, and create and share Bolt plans to expand automation across an application stack.
• Brings existing automation: works with existing scripts and plans written in YAML, PowerShell, Bash, Python, or Ruby.
• Puppet Forge reuse: content from the Puppet Forge can be pulled into Bolt workflows.
• Common maintenance tasks: patching and updating systems, troubleshooting servers, deploying applications, and stopping and restarting services.
• Cross-platform support: Linux, Windows, and macOS, with Modules, Linux, and Windows CI status badges in the repository.
• Documented onboarding: a getting started guide and self-paced training accompany the reference documentation.

Who uses it and how

• Infrastructure and operations engineers who need to run ad-hoc maintenance across a fleet without deploying agent software first.
• Teams that standardise recurring procedures such as patching, service restarts, and application deployment as shared Bolt plans.
• Mixed-estate operators working across both SSH and WinRM targets, since one tool covers Linux, Windows, and macOS.
• Existing Puppet users who want to reuse Puppet Forge content alongside their own YAML, PowerShell, Bash, Python, or Ruby scripts.
• New adopters who follow the getting started guide, the self-paced training, and the #bolt channel on the Puppet community Slack.

Getting started

Bolt installs on Linux, Windows, or macOS; installation documentation is at documentation/bolt_installing.md, and alternate installation methods and running from source code are covered in the contributing guidelines. An official getting started guide and self-paced training are available from the Bolt website.

How it compares

The provided facts name no paid products that Bolt replaces, so no head-to-head licence or cost comparison is possible here. What the facts do show is where Bolt sits: inside the Puppet ecosystem, tagged with puppet and reusing content from the Puppet Forge rather than duplicating it.

When to use it — and when not to

Bolt fits operators who already have SSH or WinRM access to their targets and want automation without installing agents; a self-hoster must install and run Bolt on a workstation and keep working credentials and network access to every target. The facts mention no hosted or managed option, so anyone wanting a fully managed service has no stated path here. The repository also carries 101 open issues against 558 stars and 224 forks, and the README stays brief, deferring operational detail to external documentation, so teams needing extensive in-repo guidance should read the linked docs before committing.