01 / Containers
Start from a Docker image.
Choose an image, set the command, and configure the container's environment and mounts. Bring private registry credentials when you need them.
Build a container task →Visual workflow builder · Public beta coming soon
Your workflows.
Every step, in view.
Connect Docker containers, server groups, AI agents, and human approvals. Build visually on your machine, then share the workflow with your team.
An example workflow
Build it step by stepFrom a container to a reviewed rollout.
- Docker
Run checks
Choose an image.
01 · Container task
Set the command. - AI agent
Review the output
Give an agent the context
02 · Agent task
and a clear task. - Human approval
Make the call
Approve, send back,
03 · Approval gate
or reject the result. - Server group
Roll out remotely
Run commands over SSH
04 · Remote command
across your servers.
Illustrative flow. Configure your own images, agents, credentials, and server targets. See how the builder works →
Workflow definitions and logs, stored locally.Powered by the Dagu workflow engine.
Different tools. One workflow.
Build the steps.
Connect the work.
Add a step, choose how it runs, and connect what comes next. Configure the workflow visually, with YAML available when you need it.
02 / Remote commands
Reach a whole server group.
Define SSH machines once, arrange them into rollout groups, and choose how many run at once and what happens when a machine fails.
Connect your servers →03 / AI agents
Give your agent a clear step.
Use a command-line agent or an API model alongside ordinary tasks. Set the prompt, provide context, and inspect the result as part of the run.
Add an AI task →04 / Human in the loop
Leave room for a decision.
Collect input with a human task or attach an approval gate to a step. Review the output, approve it, send it back, or reject it before work continues.
Add a human checkpoint →Your devices. Your team. Your organization.
Build once.
Bring your team along.
Move a project between your devices or share it with teammates across your organization. Workflow definitions and reusable configuration travel together. Each device supplies its own credentials and chooses what runs.
Share a workflow with your team →Take the project to another device
Export a project and import it on another Runstead installation. Bring its workflows, agents, server groups, and defaults together.
Give teammates a shared starting point
Hand over a repeatable workflow instead of a set of instructions. Imported schedules start paused, so each person decides when to run them.
Connect to a shared Runstead host
Let team members use MCP or REST clients with separate API keys and role permissions. Configure the network access for your environment.
From building to running
Make it repeatable.
Keep it understandable.
Run on demand or set a schedule. Follow each step, inspect its output, and find what needs attention. Add inputs, retries, and notifications as the workflow grows.
Your machine. Your workspace.
Keep your work
close to home.
Runstead hosts the workflow engine on your machine. Your workflow definitions, execution logs, and backups are stored locally. Tasks and connected clients can send data to the services you choose.
Understand background operation →Keep working in the background
Keep Runstead running while you work. On macOS, closing the window leaves it in the menu bar. Scheduled work needs an awake machine with your user signed in.
Connect on your terms
Use scoped API keys for MCP and REST access. Remote access requires your own secure network setup.
Start without an account
The free plan holds one project. A Dagu Cloud account is only needed for Pro.
Start small. Make it useful.