Managed databases and apps. Minus the surprise bill.
Selfhost.dev runs fully-managed PostgreSQL®, MySQL®, Redis® and ClickHouse® on AWS, plus apps from GitHub and 26 one-click configurations across 23 apps in the console, including Supabase, n8n, WordPress, GitLab and Open WebUI + Ollama. The platform also has 189 AI/MCP tools. One account, one bill and you pause to $0 anytime.
Spin up a project and ship your app
A real engineer plans and runs your move, at no cost. Get a free audit
Your AI DevOps engineer, in your editor
Pick how it runs.
The engine is the easy part.
Four managed engines, two extensions, one project server. One account, one bill, one AI control plane. The only decision that changes what you get is the first one.
selfhost=# SELECT engine, capability, price selfhost-# FROM selfhost.catalog selfhost-# ORDER BY hosting_model;
A dedicated instance on AWS, in our account or your own. Automated backups, Multi-AZ with automatic failover, forking, scaling, live metrics and alerts. Worldwide regions. PITR and connection pooling on the SQL engines.
PostgreSQL 16, 17, 18
Restore to any second. PgBouncer pooling on port 6432, Multi-AZ failover, forking, parameter tuning, users and roles and BYOC into your own AWS account.
pgvector Vector database
A switch on the instance above, not another database. Embeddings and similarity search for RAG, in the same Postgres, on the same connection string and the same bill.
TimescaleDB Time-series database
The same table, chunked by time. Hypertables, native columnar compression and continuous aggregates for metrics, events and IoT. No second system to run.
ClickHouse columnar
Reads one column across billions of rows. Real-time analytics on HTTP 8123 and native 9000, replicas and automatic failover through an embedded Keeper quorum. About 2.6x less than ClickHouse Cloud for the same shape.
Redis 7.0, 7.2
0.88 ms at the 99th percentile. Measured at about 230k operations per second. Cache, queues and sessions, with snapshots, replicas, resize and Multi-AZ failover. The 2.1x is the same r7g.large class in ap-south-1, against ElastiCache at AWS list price (estimated).
MySQL 8.0, 8.4
PITR to the minute. ProxySQL pooling, Multi-AZ failover, forking, users and roles and BYOC into your own AWS account.
One click on your own dedicated project server, next to the app that uses it. Simpler on purpose: create, delete, live metrics and an activity log. No PITR, no pooling, no version choice. 2 to 48 vCPU, from $0.02/hr and stopping the server stops the bill.
Apps, APIs and their databases
Deploy from GitHub with auto-deploy on push. PR previews, custom domains and SSL, env vars, streaming logs, rollback. One-click PostgreSQL, MySQL, Redis or MongoDB alongside it.
26 configurations across 23 one-click apps
From Supabase and n8n to GitLab and Open WebUI + Ollama. Each installs inside a project, on the server you already pay for. Browse the full console catalog and pick the stack that matches the job.
We put each workload on the cloud that does it best and handle the ops. You just get their uptime and performance.
Manage core operations
with AI.
189 MCP tools across 31 modules connect managed databases and core Projects operations to Claude, Cursor or any MCP client. For one-click deployments, MCP currently supports four configurations: Supabase, n8n (SQLite), n8n (Postgres) and Twenty. The broader 26-configuration catalog remains available in the console.
Architect your ecosystem.
Code, databases and templates on one server.
A transparent PaaS: name it and we provision a real dedicated server for you, from 2 to 48 vCPU, kept in full view. Then drop in a GitHub repo, a database or one of 26 one-click configurations across 23 apps, each with custom domains and SSL handled for you.
Deploy from GitHub, auto-deploy on push, PR preview environments, rollbacks.
Any database: PostgreSQL, Redis, MySQL, MongoDB, one click each.
One-click apps: browse all 26 configurations across 23 apps, from Supabase and n8n to GitLab and Open WebUI + Ollama.
Custom domains + SSL, add a domain, verify DNS, go live on HTTPS.
Services on this server
One-click apps
26 ready-made configurations across 23 apps
Supabase
8 GB RAM minimum
n8n
4 GB RAM minimum
n8n (Postgres)
Bundled Postgres variant
Twenty
8 GB RAM minimum
WordPress
4 GB RAM minimum
GitLab
8 GB RAM minimum
Elasticsearch
8 GB RAM minimum
Immich
8 GB RAM minimum
Nextcloud
8 GB RAM minimum
Jenkins
4 GB RAM minimum
Grafana
4 GB RAM minimum
Grafana (Postgres)
Bundled Postgres variant
Elasticsearch + Kibana
Kibana UI included
Chatwoot
8 GB RAM minimum
Mautic
8 GB RAM minimum
Strapi
4 GB RAM minimum
Metabase
4 GB RAM minimum
Odoo
8 GB RAM minimum
Paperless-ngx
8 GB RAM minimum
Stalwart
4 GB RAM minimum
MinIO
4 GB RAM minimum
Open WebUI + Ollama
8 GB RAM minimum
LiteLLM
4 GB RAM minimum
AnythingLLM
4 GB RAM minimum
EspoCRM
8 GB RAM minimum
Krayin
8 GB RAM minimum
Production Postgres, MySQL, Redis and ClickHouse.
In our cloud or your own AWS account.
Production-grade reliability, encryption at rest, backups, PITR and Multi-AZ, without hiring a DBA. Explore managed database hosting on AWS.
Live in under 2 minutes
Pick your instance and engine (PostgreSQL, MySQL, Redis or ClickHouse), hit deploy. VPC, security groups, backups, monitoring, alerts and Multi-AZ failover, all configured automatically on AWS. On PostgreSQL and MySQL, PITR restores to any second, with connection pooling built in.
Deploy a databaseYour AWS account. Our management.
Run managed PostgreSQL, MySQL, Redis or ClickHouse inside your own AWS account. You own the infrastructure; we handle provisioning, monitoring, backups, scaling and alerts. Your data never leaves your account and there's no lock-in.
No sales calls. No waiting.
Launch one with MCP in under 30 seconds.
Deploy across AWS, worldwide
Run your database close to your users on AWS, for lower latency and higher reliability, wherever they are.
Real-time monitoring, PITR and auto-pause
Continuous visibility into performance, point-in-time recovery to any second, database users with rotating credentials and scheduled start/stop to pause idle databases. Alerts reach you before issues escalate.
Your data: encrypted, isolated,
and always under your control.
Security is not a separate tier here. Every managed instance ships locked down by default, encrypted end to end and yours to export or run in your own cloud.
See how we secure and run your dataEncrypted in transit and at rest
TLS on the wire, AWS KMS on disk. Snapshots and backups are encrypted too.
Single-tenant, dedicated instance
Your own dedicated instance, not a shared multi-tenant database pool.
Private by default
IP whitelisting and delete protection are on. Public access only if you turn it on.
Multi-AZ, PITR and backups
Automatic failover on every engine, point-in-time restore on PostgreSQL and MySQL, automated backups on all.
Your data stays yours (BYOC)
Run it in your own AWS account. Standard engines, nothing proprietary, yours to keep.
Fast and proven in the open.
Full durability on, no special tuning.
Real pgbench numbers on AWS, with production durability settings and zero errors across every run. Not a cherry-picked demo. See the same workload priced against AWS RDS.
Pay only for what you run.
Prepaid credits, no tiers and no subscriptions. Everything is billed by the hour against your balance and pauses at a zero balance, so you never get a surprise invoice: no mystery IOPS or egress bill at month-end.
A dedicated PostgreSQL, MySQL, Redis or ClickHouse instance on AWS, run for you and metered by the hour: from about $0.007/hr or about $0.033/hr for ClickHouse, whose smallest instance is a 4 GB box. Multi-AZ on all four, PITR and pooling on the SQL engines.
A dedicated 2 vCPU / 4 GB RAM / 40 GB SSD server with 1 TB bandwidth, Hetzner's entry tier in the EU. Deploy apps from GitHub with the databases they need. Pause it and pay nothing.
The cost of self-hosting.
The control of managed. No tiers.
We spent a decade in startups watching teams burn time managing five different providers for one product. A database host here, an app host there, a cache somewhere else. Different dashboards, different bills and half the features they paid for, they never used.
Simple things like creating a database user from a UI or managing your infra from your code editor required a dedicated DevOps person. And if you tried to self-host, you spent your nights patching servers and debugging Traefik configs instead of building.
Selfhost.dev is what we wanted: everything your product needs on one platform. A managed PostgreSQL for $5/mo and you get unlimited alert rules, unlimited org users, all the features, no gating. The cost of self-hosting with the control of managed.
One platform, not five
PostgreSQL, Redis, MySQL, ClickHouse, MongoDB, app hosting and one-click apps. One account, one bill, one dashboard. No hopping between providers.
Everything included
A $5/mo database gets you unlimited alert rules, unlimited org users and every feature its engine supports, PITR and pooling included on the SQL engines. No tiers, no add-ons, no gating.
AI-native from day one
189 MCP tools for managed databases and core Projects operations. One-click deployment tools currently cover Supabase, two n8n configurations and Twenty.
Complete operational visibility
across databases and projects.
Performance monitoring
Track CPU, memory, disk I/O, latency and throughput in real time to understand how your databases and services perform under load.
ExploreOperational status
Track everything across your databases and projects: provisioning, deploys, scaling, backups, domain and config changes and user actions, from one timeline.
ExploreShip on every push
One-click and auto-deploys from GitHub, with build logs, preview environments and instant rollbacks.
Custom domains, wired up
Add a domain and SSL is handled for you. Pipe deploy and health events to Slack or any webhook.
System health overview
Instant visibility into the health of your databases, servers and deployments so you can catch issues before they reach users.
| Instance | Status | Latency |
|---|---|---|
| prod-db-01 | STABLE | 1.2 ms |
| replica-prod | STABLE | 0.8 ms |
| app-uat-02 | PEAK | 4.5 ms |
| app-uat-01 | PEAK | 4.1 ms |
| web-prod-db | STABLE | 2.3 ms |
| web-uat-db | HIGH | 9.3 ms |
Smart insight for you
Infrastructure health has improved by 12% compared to last week.
Also, no critical errors detected in the past 27 hours.
This isn't just a screenshot. It's a fully interactive demo of the console and you can drive it right now. No signup.
Open the live demo
Find out why teams
choose Selfhost.dev
Already trusted by early teams and developers to run real apps and databases in production.
Our RDS bill kept climbing with IOPS and egress charges nobody could fully explain. On Selfhost.dev the same Postgres runs for a fraction and prepaid credits mean I watch the burn rate instead of dreading the invoice.
I run our whole database from Cursor now: provisioning, backups, scaling, all in plain English. Spun up a production Postgres for our ML pipeline in 3 minutes without leaving my editor.
Our team spent 15-20% of every sprint on database ops: scaling, backups, incident response. That's close to zero now. And it runs in our own cloud, so if Selfhost.dev vanished tomorrow, our data would still be there. That's rare.
Frequently asked questions
Is Selfhost.dev the same as self-hosting?
What is Selfhost.dev?
What can I deploy on a project?
How does pricing work?
Is the free migration really free?
Which databases can I run?
What is BYOC?
How many MCP tools does Selfhost.dev have?
How is Selfhost.dev different from Railway or Render?
How do I get started?
Why self-host when
Selfhost.dev can run it for you?
PostgreSQL, Redis, MySQL, ClickHouse, MongoDB, apps, APIs and one-click apps. Everything your product needs on one platform, run for you, with core operations available through AI. Pay only for what you run.