Memory is the constraint
Idle workspaces and previews stop to free memory.
Hosting
DevBench runs on a host you control and deploys to the cloud you already have: AWS, a server over SSH, or DevBench compute for previews.
One container per workspace, on a host you control.
Previews run on DevBench compute, free and throwaway.
Where should it run?
DevBench compute
Free previews for the whole team to test in.
What you need
What DevBench does
Who may press it
Anyone can deploy a preview.
Agents can deploy previews, not environments.
$ devbench deploy
Deploy this workspace's repos as live test containers. The same as the Deploy button.
DevBench runs on a machine you control, with Docker. Your code leaves it only when you deploy.
Outside
Your host · A machine with Docker
Leaves the host
Control plane · Runs on your host
Runs on your host in Docker. The chat updates live.
Idle workspaces and previews stop to free memory.
Your owner creates the organization and invites the team.
You choose the model. Nothing else depends on it.
Every repo runs as a container you can click through. Each deploy replaces the last.
Replaced by the next deploy. Stopped when idle.
Preview runs on DevBench compute. Environments has one row per branch.
Your own pipeline deploys it. DevBench keeps the record.
Free previews the whole team can open and test in.
A site or server on AWS, or a machine you reach over SSH.
Missing permissions show up first.
It updates the existing site or server, and never makes a new one.
Stop marks the deployment stopped and deletes nothing.
An unknown target fails with an error. It never falls back to DevBench compute.
Your team writes an update script for each environment. Without one, the deploy is refused.
set -euo pipefail # DevBench adds this line cd /srv/app git fetch origin "$DEVBENCH_BRANCH" git checkout -q "$DEVBENCH_BRANCH" git pull --ff-only npm ci && npm run build sudo systemctl restart app # yours: systemd, pm2, compose, make…
The host key is checked on every connection.
A failed step stops the script before any restart.
Commands run as you wrote them.
The branch and environment are passed in, so one script serves them all.
AWS
Credentials say who DevBench is in an account. Switches say what it may build there.
Use an IAM user's keys, or a role DevBench can assume.
Checks the permissions without changing anything.
The result is a warning, never a pass.
As a policy you can paste.
Each AWS connection is checked again on a schedule. A failure reaches the owners.
Lists what already exists in the account. Each item can be linked to the part of the app it serves.
$ devbench cloud can
What this workspace may do in the cloud, and how far.
$ devbench cloud list
Deployments that already exist in your cloud accounts.
$ devbench cloud link <id> <repo>[/<dir>]
Record which part a resource serves.
$ devbench cloud call <service> <Operation>
One AWS call, made through DevBench.
DevBench deploys to it the same way it deploys to an instance it created.
Any machine you can reach over SSH
Connecting asks forHost, port, user, a key or a password
Supplies machines
Connecting asks forA workspace API token
Supplies machines
Connecting asks forA personal access token
Supplies machines
Connecting asks forAn API key
A Supabase workspace
Connecting asks forWorkspace URL, service role key, anon key
Atlas or your own mongod
Connecting asks forA connection string
Each capability is checked by running its tool. A missing package shows the command that installs it.
The vendor accounts supply machines. You choose a size and region at deploy time.
Apps
An App groups the repos of one application.
Each part has its own recipe. Write it by hand, or let DevBench propose one. Nothing changes until you apply it.
Each App owns its databases and seed SQL. Schema changes go in a migration.
$ devbench deploy
Deploy this workspace's repos as live test containers. The same as the Deploy button.
$ devbench deploy config list
Every deployable part of this workspace and where it goes.
$ devbench deploy config generate
Read a repo and propose how each part builds and runs.
$ devbench deploy config apply <id>
Apply a proposal.
$ devbench deploy config set <repo>[/<dir>] --start … --build …
Create or edit one part's recipe.
$ devbench deploy seed set --file <path>
Replace the seed data.
Set up with you
We set up your factory with you and ship one project from your roadmap.