Deploy like Heroku.
Pay like a VPS.
Hakopod installs K3s, PostgreSQL and a dashboard on one Linux server. You deploy from Git, an image or a docker-compose file, and each app gets HTTPS, private networking and a plan you review before it ships. The server is your main bill.
curl -fsSL https://hakopod.com/scripts/installer.sh | sudo shOr create a Hakopod Cloud account (beta) →
Apache-2.0 · 80+ releases since September 2026 · 68 templates · 4 GiB RAM · AMD64 and ARM64 · Alpha: read what's tested
Watch a one-minute tour of Hakopod.
An app deployed, its services connected privately, and its logs and history in the dashboard.
English captions included
Read film description
The current dashboard (v0.1.0-alpha.89) with a demo shop app: a service's live usage, its TOML, a recorded deployment and the application overview.
Apps, databases, cron, GPUs and CI.
Anything that runs in a container, plus the services around it. One TOML file describes an app, and the dashboard edits the same file.
- Web apps and APIs Git, image or Compose
- Managed databases 8 engines
- Workers, cron and one-off jobs Next to the app
- Serverless functions Scale to zero
- Open models with vLLM NVIDIA GPUs, beta
- GitHub Actions runners Pro, beta
schema_version = 1
name = "shop"
[services.web]
image = "ghcr.io/example/shop-web:1.4"
port = 8080
public = true
replicas = 2
[services.worker]
image = "ghcr.io/example/shop-worker:1.4"
[services.report]
image = "ghcr.io/example/shop-reports:1.4"
command = ["node", "report.js"]
[services.report.job.schedule]
cron = "0 * * * *"
timezone = "UTC"
Let Claude Code fix the 502.
Hakopod ships an MCP server and a CLI. Claude Code, Cursor or any MCP client can read logs, find what broke, plan a fix and deploy it.
{
"mcpServers": {
"hakopod": {
"command": "hakopod",
"args": ["mcp", "--project", "shop",
"--environment", "production",
"--allow-deploy"]
}
}
}resource "hakopod_project" "shop" {
name = "shop"
environments = ["production"]
}
resource "hakopod_application" "web" {
project = hakopod_project.shop.name
environment = "production"
config = file("hakopod.toml")
}
Reading apps, logs and the live commit is on by default. Deploys and
rollbacks need --allow-deploy, database queries --allow-sql, pod shells --allow-exec, and most of the 329 API
operations --allow-write. Keys belong to one project
and environment, and they expire. Login, MFA and consent stay with
people. The GitHub Action is hakopod/deploy@v1.
GitHub Actions runners at Spot prices.
Change one runs-on label. Each job gets a fresh runner that
is used once.
jobs:
test:
# was: runs-on: ubuntu-latest
runs-on: [hakopod]
steps:
- uses: actions/checkout@v4
- run: npm test| Runner | Size | 6 min | 20 min | 60 min |
|---|---|---|---|---|
| Hakopod on AWS Spot, c7i or c6i.xlarge | 4 vCPU, 8 GB | ~$0.006 | ~$0.020 | ~$0.058 |
| Hakopod on AWS Spot, m7i.xlarge | 4 vCPU, 16 GB | ~$0.011 | ~$0.033 | ~$0.099 |
| GitHub-hosted Linux, 2-core | 2 vCPU | $0.036 | $0.12 | $0.36 |
| GitHub-hosted Linux, 4-core | 4 vCPU | $0.072 | $0.24 | $0.72 |
- 1,000 6-minute jobs a month: about $6–11 on Spot VMs, against $72 on GitHub's 4-core runner and $24 on its 2-core runner after the free minutes.
- 1,000 20-minute jobs a month: about $20–33 on Spot VMs, against $240 on GitHub's 4-core runner and $108 on its 2-core runner after the free minutes.
Estimates for us-east-1, November 2026, from public Spot averages (about $0.05–0.09 an hour) and GitHub's published per-minute rates. A VM bills about half a minute more than its job, for boot and shutdown, and the figures include its disk and public IPv4. Spot prices change by hour and zone. Not included: data sent out of AWS, such as pushing an image to a registry (about $0.09 per GB), and the Hakopod server and licence, which cost the same at any job count.
- Pods on your own servers gVisor sandbox
- AWS EC2 Spot Released
- Google Cloud Compute Engine Preview
- Microsoft Azure Preview
A runner is either a gVisor-sandboxed pod on your own servers or a single-use VM in your own cloud account. VMs run Docker natively. The VM is deleted when the job ends, and a job lost to a Spot interruption reruns on its own. Hakopod stores no cloud keys: it assumes an IAM role on AWS, impersonates a service account on Google Cloud and uses its own identity on Azure. Set it up under Settings → Actions VM capacity, with the IAM, gcloud and az steps ready to copy and a Test button.
Read about Managed Actions ↗
Private by default. Free to self-host.
The code is on GitHub under Apache-2.0. Self-hosting costs nothing; you pay for the server.
Services start private, and network rules block traffic between apps until you allow it. Secrets are encrypted and masked in logs. CI and agent keys belong to one project and environment, and they expire. Every hour, each running app's image and runtime are checked against a signed advisory feed and OSV.dev.
Read how it is secured ↗ Read the source ↗- Self-hosted FreeThe full platform, teams and fixed project roles.$0
- Self-hosted ProAudit history, SSO, Slack and Managed Actions runners (beta). Or $29/month.$99/ year
- Cloud StarterA managed dashboard for your own server. Beta; paid checkout is not open yet.$5/ month
Before you install.
curl -fsSL https://hakopod.com/scripts/installer.sh | sudo sh01 Do I need to know Kubernetes to use Hakopod?
No. Hakopod installs K3s and creates the Kubernetes objects for every app, including ingress, certificates and network rules. You deploy from the dashboard, a short TOML file, a docker-compose file, the CLI or an AI agent, and never write YAML or run kubectl. You still own the server itself, its firewall and K3s upgrades.
02 How is Hakopod different from Coolify and Dokploy?
Coolify and Dokploy run your apps on Docker; Hakopod runs them on K3s, a lightweight Kubernetes. That needs more memory, at least 4 GiB RAM, but a second server joins as a worker and rolling updates come from Kubernetes. Hakopod also shows every deploy as a plan before it applies. The compare pages go through each platform in detail.
03 What do I need to run Hakopod?
A fresh Ubuntu 24.04/26.04 or Debian 12/13 server, AMD64 or ARM64, with at least 4 GiB RAM and 30 GiB of free disk. Run the one-line installer with sudo; it sets up K3s, PostgreSQL and the dashboard. The installation guide covers DNS, firewall and provider-specific steps.
04 Is Hakopod ready for production?
Not yet. Hakopod is an alpha release, and the control plane and ingress have no high availability yet. Read the installation guide and release notes for what has been tested before moving anything important onto it.