PaaS markup on their servers
Heroku, Railway, and Render run your app on infra they control.
Your bill grows faster than your product.
Deploy on your infrastructure
Git push deploys. On servers you already own.Ship at PaaS speed.Keep your cloud bill.
Your apps. Not platform markup.
- Your worker nodes
- No vendor lock-in
- Git to live URL
app.dflow.sh / Release map
Deploy orchestration
Git push
GitHub repo
Control plane
Your project dashboard
Worker node
Your VPS / cloud
Git push
GitHub repo
Control plane
Your project dashboard
Worker node
Your VPS / cloud
On worker node · apps & databases
api-gateway
Running2m ago
postgres-main
Backed up842 MB
preview-pr-142
Readyauto-sleep
Latest deploy · api · v2.1.0
api.acme.dflow.sh
worker-01 · 3m 08s · 0 errors
The markup trap
You should not pay extra for compute you already own.
PaaS platforms host on their infra and pass the markup to you. DIY DevOps burns weeks before the first deploy. Every path makes you trade speed or your bill.
DIY before the first deploy
Kubernetes, Terraform, and custom CI before anything is live.
One engineer becomes the bottleneck.
Split tools, slow fixes
Hosting here, databases there, logs in a third place.
Incidents take longer when the stack is scattered.
Same setup tax every project
Agencies rebuild hosting from scratch per client.
Repeat delivery work eats margin.
Your stack
Whatever you build, it ships on dFlow.
Push your repo. We detect the framework, build on your worker node, and hand back a preview URL. If it runs in Docker, it runs here.
Git pushYour worker nodesPreview URLs
Frameworks and runtimes
Next.js, Python, Go, Java, and more, detected from your repo on push.
- Next.js
- Vue
- Node.js
- Python
- Go
- Spring Boot
- Java
- Ruby
- PHP
- .NET
- Scala
Why dFlow
Use the servers you already pay for.
Cloud credits, existing instances, your VPS fleet. dFlow turns them into a developer platform, without a platform team and without a platform markup.
Every infra dollar stays on your bill.
Pay your cloud provider directly. No per-seat markup on compute you already own.
cloud.bill / your-account
Worker node (your VPS)$12/mo
Postgres (same node)included
Managed worker (optional)usage-based
Platform markup on computenone
Your servers, your data
Workloads, databases, and secrets stay on infrastructure you control. Disconnect anytime.
Compute that fits
Right-size worker nodes, spin up preview envs on demand, auto-sleep idle staging when configured.
Your control plane
Run dFlow on infrastructure you operate. Same deploy model when you add managed workers or enterprise support.
Deploy your way
Git push, dashboard, or MCP from your AI editor: one control plane, the same path to a live URL on your worker nodes.
Control plane ops
Ship, store, connect, and grow on your nodes.
Deploy pipelines, databases, domains, and capacity from one dashboard or MCP prompt, always on worker nodes you connect, not a shared pool.
DEPLOY
Deploy from git or MCP
Push code or prompt your agent. dFlow detects the stack, builds on your worker node, and returns a preview URL.
worker-01 · api-service
09:14:01 Push to main · 8 env vars loaded
09:14:22 Next.js 15 build started
Preview URL ready
DATA
Databases beside your app
Postgres, MySQL, MongoDB, or Redis on the same worker node. Persistent volumes and a connection string from the dashboard.
worker-01 · postgres-main
10:02:00 PostgreSQL 16 provisioned
10:02:12 20 GB volume attached · linked to api-service
Connection ready
NET
Domains and private mesh
Custom domains with automatic SSL, plus Tailscale so services talk privately on your infrastructure.
acme-corp · production
11:02:10 Let's Encrypt cert issued for api.acme.dev
11:02:15 Tailscale mesh linked · api, postgres, redis
HTTPS live
SCALE
Scale when you need it
Raise replica counts, adjust CPU and memory limits, or register another worker node from the dashboard.
worker-01 · web process
18:45:00 Replicas updated 3 → 5
18:45:18 Rolling deploy finished · all instances healthy
5 replicas live
From push to production
One git push. Your worker nodes take it from there.
Build, ship, deploy, and watch on the servers you already pay for.
01SERVER
You connect
Register AWS, DigitalOcean, on-prem, or any VPS. dFlow installs a scoped agent. Your credentials stay yours.
$ dflow server add worker-01
Agent installed · scoped access only
Worker node ready.
Platform
Your applications workspace, not a generic cloud console.
Browse applications, worker nodes, and integrations from the same dashboard you use to approve GitHub deploys.
app.dflow.sh / acme-corp / applications
ApplicationsNewWorker NodesIntegrations
Applications
Manage your applications, environments, and usage limits.
api-platform
activeCore API and background workers
GitHub · acme/api-platform
productionstaging
4 servicesworker-01
storefront
activeNext.js storefront and checkout flow
production
2 servicesworker-01
Pending deployment
feat: rate limit middleware
pending approvala3f91c2mainapi-gateway · production
DiscardApprove
Deploy complete
a3f91c2 live on worker-01 · zero downtime rollout
GitHub connected
acme/api-platform pushes create deployments automatically
Approval required
Review pending deploy before it reaches production
How it works
Zero to production in three steps
Self-serve from a browser or your AI editor. No sales call required.
Connect a worker node
A few minutes in the browser. Add AWS, any VPS, or on-prem. dFlow only gets the access it needs to ship your apps.
[access scope · least privilege]
- Run services on tagged worker nodes
- Read deploy logs and health status
- Store build artifacts in your storage
- Create cloud users or change billing
- Touch resources outside your project
Push your code
Git push, or wire the dFlow MCP in Cursor and ask your agent to deploy. You get a preview URL you can share.
$ git push origin main
Built on your worker node
Live at api.preview.acme.dev
Ship and iterate
Domains, databases, rollbacks, and team permissions from one place. Deploy workflows watch the stack from there.
Deploy workflow
api-gateway rolled out · preview URL ready
Compare
How dFlow compares
PaaS-style shipping on your servers, without the markup or lock-in.
| Dimension | DIY DevOps | Heroku / Railway | dFlowRecommended |
|---|---|---|---|
| Time to first deploy | Weeks | Minutes | Minutes |
| Runs on your servers | Yes | No | Yes |
| You pay your cloud bill directly | Yes | No | Yes |
| SOC 2 / HIPAA compliance scope | Your cloud, your BAA | Add-ons · vendor scope | Your provider, your BAA |
| Data residency control | You pick account & region | Vendor regions only | AWS, DO, Azure, or your VPS |
| Deploy from AI editor (MCP) | No | No | Yes |
| Apps + databases unified | Manual | Partial | Yes |
| Preview environments | Custom build | Varies | Yes |
| Platform fee at scale | Engineer salary | High markup | Usage-based · custom enterprise |
| Vendor lock-in | Low | High | Low |
winpartialmissing
Compliance follows your provider
SOC 2, HIPAA, ISO, and regional controls inherit from whichever worker-node provider you connect, not from a shared multi-tenant PaaS pool.
- AWS
- DigitalOcean
- Azure
- Hetzner
- Any VPS
Pricing
Managed Cloud adds usage-based worker capacity. Enterprise includes SSO, hybrid deployment, and custom commercial terms.
Enterprise
For teams that need SSO, flexible deployment, and support matched to how you ship. Tell us about your stack and environments-we will recommend the right setup and commercial terms.
- SAML/OIDC single sign-on (SSO)
- Custom roles and team permissions
- Self-hosted, managed, or hybrid deployment
- Guided onboarding and enablement
- Priority support with SLAs
- Volume pricing for larger teams
Custom pricing
Let's talk
Plans are scoped to your deployment model, team size, and support needs-not fixed public tiers.
FAQ
Frequently asked questions
Every answer comes back to the same point: ship on your servers, pay your cloud bill, skip the platform markup.
Can't find what you're looking for? Contact our customer support team