Your team self-serves.
Developers wake, clean, and manage their own boxes in one click. No tickets, no 7am SSH, no devops interrupt.
Illustrative. Swap for a product screenshot. Download the appUplink by Deployport
Stop babysitting
your servers.
Your team wakes and cleans up their own boxes and idle VMs shut off on their own, so your platform engineer stops babysitting environments and gets back to the work only they can do. Enroll today, no migration.
curl -fsSL https://get.deployport.com | sh Then enroll the box with a key from the console. Free for your first 3 hosts. No credit card.
Who it's for
If your platform team is one or two people, this is for you.
Uplink is built for growing engineering orgs, roughly 20 to 100 developers, whose entire platform and devops function is one or two people, buried in dev-environment tickets while the infrastructure that actually matters waits. If that is you, this was built for you.
How it works
From two commands to a lower bill, in three steps.
- 01
Install.
Two commands on the Ubuntu servers you already run: install the CLI, then enroll the box with a one-time key from the console. After that the agent keeps its own device identity, so it survives reboots and months offline. No migration, no new platform.
- 02
It tells you the truth.
Disk, memory, hardware. When something is wrong it files an issue in Slack or email that stays open until it is actually fixed, with the fix attached.
- 03
It saves you money.
Schedule VMs off after hours, stop them when they go idle (without killing a running build), and let your team wake them with one click.
Provisioning
Your developers stop asking you for machines.
You decide what shapes are allowed. They launch inside that, on your cloud account, whenever they need one. Nobody files a ticket and nobody gets a bill they did not expect.
You set the bounds once
- Regions
- Only the ones you allow
- Size
- A ceiling on vCPU and memory
- Architecture
- x86, ARM, or both
- Expensive shapes
- Bare metal and GPU off by default
They launch inside them
A developer picks from what you permitted and gets a machine in under a minute, already enrolled, already under your idle and disk policies.
Anything outside the bounds is refused, by name, with the reason. Not hidden in a menu they can turn back on.
The math, two numbers not one
Pays for itself twice: the cloud bill, and the engineer.
The obvious return is the cloud bill: a VM running around the clock costs roughly $150/month; shut it off when it is idle and you cut about a third of the whole bill, thousands a month across a team. That is the number that clears procurement. The bigger return is your scarce platform engineer: every hour they spend babysitting dev environments, fielding "turn my box on" tickets, and hand-cleaning disks is an hour not spent on critical infrastructure. Uplink gives those hours back.
Platform-engineer time reclaimed 0 hrs/mo worth about $0/mo of engineering time, the reason to buy
Idle cloud spend removed $0 per month
Estimated Uplink cost $0 per month, at $7 per host, month to month
Net cloud savings $0 the CFO permission slip
Roughly $200/month of Uplink against $2,000 to $3,000/month of idle spend removed, plus a platform engineer freed from dev-env toil.
What you get
Your platform engineer stops babysitting dev environments.
That is the outcome. Here is how it happens.
The boxes behave.
Predicts a disk filling up ("/ is full in about 18 hours"), names the hog (Docker build cache), catches the kernel out-of-memory kill nobody sees, and tells you when a box goes dark.
Illustrative. Swap for a product screenshot.The bill drops.
Idle VMs shut off on their own, without killing a running build. Developer VMs get about 70% cheaper, which lands as roughly a third off a typical whole cloud bill.
Illustrative. Swap for a product screenshot.What Uplink is not
Not a platform. It rides the VMs you already have.
Uplink does not provision environments, replace your workflow, or ask you to move anything.
vs a CDE (Coder, Codespaces)
They create and own the environment, so adopting one is a migration. Uplink installs in two commands on what you already run.
vs SSM or remote shells
Those give your team root with extra steps. Uplink gives them a fixed set of buttons, dry-run first, every action audited.
vs metrics tools (Netdata, Datadog)
They hand you graphs and homework. Uplink ships a verdict with the fix attached.
Free your devops, without the friendly fire
Free your platform team, without breaking everyone else's flow.
The usual fix is a cron or a fixed schedule that turns idle boxes off. It frees your devops person and then it steps on the team: it kills a running build, sleeps a box someone is still using, and leaves developers filing a ticket to get their environment back. So teams either never do it, or do it once and rip it out.
Uplink is the only version that frees your devops and keeps DevEx intact:
- CPU and network aware. It will not stop a box running a build or actually in use.
- Warns before it sleeps a box, with one-click HOLD to keep it alive.
- Developers wake their own boxes in one click. No ticket, no devops interrupt.
Your platform engineer stops babysitting dev environments, and your developers never feel it.
Trust and security
Keyless by design. Your cloud credentials never leave your account.
You grant Uplink a scoped, least-privilege role in your own AWS or GCP account, and revoke it in one click. We store no long-lived secret, ever. The agent is read-only by default: every action is opt-in, dry-run first, allowlisted, and audited. The agent is open source (AGPL).
- No stored secrets, ever.
- Least privilege, in your own IAM, revocable by you.
- Read-only by default. Actions are dry-run first.
- Open-source agent (AGPL).
- We never sell your data.
Free
$0
3 hosts, 3 users, no credit card. Genuinely useful forever.
- Every diagnostic
- Manual power control
- 3 hosts and 3 users
Team
$5 / host / mo
Billed annually. Seats included.
- Auto-off and schedules
- Idle auto-stop
- Self-serve power
- Fleet view
- Savings reports
Enterprise
Custom
Controls and compliance for larger fleets.
- Enrollment audit
- Enforced policies
- SCIM
- Audit streaming
- Self-hosted control plane
- SLA
From the founder
We were spending about 9k a month on dev environments, and my one devops engineer was constantly getting distracted by vm-related tickets. At some point we introduced a weekend and nightly schedule to turn those machines off, and it was so inflexible it hurt more than it helped. So I built the version I actually wanted, one that frees up the devops person without the rest of the team feeling it.
Johan, Deployport