Fluid enables SREs, Platform Engineers, and DevOps Engineers to fix
multiple issues at once, expand their triage and enable more uptime
across their production environment.
Fluid works in four phases.
> read
Debug production VMs with Fluid's read-only mode. Let Fluid
investigate an issue by querying log files, reading systemctl changes,
or accessing config.
> edit
Edit VM sandboxes with Fluid's edit mode. After Fluid has some context
on the system, it will create a sandbox by cloning the VM. Fluid will
then make changes, edit files, and iterate within the sandbox until
the issue is resolved.
> ansible
Once Fluid has fixed the issue on the sandbox, it will begin to create
an Ansible playbook to reconstruct the fix on production.
> cleanup
After the changes are made, open sandboxes will be deleted when the
Fluid CLI is closed or you can ask Fluid to delete them manually.
Installation
This will install the $fluid.sh terminal agent meant
to be installed on your local workstation.
If you are instead looking for the $fluid.sh agent api, to control
thousands of concurrent agents/sandboxes, navigate to the readme.
$ go install
github.com/aspectrr/fluid.sh/fluid/cmd/fluid@latest
$ curl -fsSL
https://fluid.sh/install.sh | bash
$ fluid
Built for where you already work
[~]
Sandbox Isolation
Clone VMs instantly. Test changes in isolation before touching production.
ls
Context-Aware
Fluid explores your host first - OS, packages, CLI tools - then adapts.
>>>
Full Audit Trail
Every command logged. Every change tracked. Review before production.
.yaml
Ansible Playbooks
Auto-generates playbooks from sandbox work. Reproducible infrastructure.