Record of transfer
Selfhost.dev, a managed-database and app-deployment platform
Operated by Swatantra Cloud Systems Pvt Ltd, Pune · Building since 2025
Prepared for: the undersigned developer · Recorded: 2026-08-16
The managed home for
your whole backend.
This page is written as a deed, because that is what it describes: a record of what belongs to you when infrastructure is run the way we run it.
In the matter of: managed PostgreSQL, MySQL, Redis, and ClickHouse on AWS · Projects, a managed server for your apps and databases · 161 MCP tools · prepaid credits
Part I The lease you already signed
You have read this document before. It was set in a friendly sans-serif, it scrolled forever, and you agreed to it without reading. Everyone did. The modern cloud runs on a lease, and these are its actual terms.
The lease's central term is never printed. It is this: the machine is ours. You may rent outcomes.
Part II Recitals
Whereas our founder's first server was a cheap shared box that ran everything: client sites, four WordPresses, cron jobs held together with hope. One login. One bill smaller than lunch. Every byte of it understood.
Whereas that box could be drawn on a napkin, because it had been drawn on a napkin.
Whereas the industry's growing-up story required scattering it: a queue here, a function there, a database behind someone's premium, until understanding your own stack became a full-time job nobody was hired to do.
Whereas the old box was honest about both sides of itself. What it got right: comprehension, consolidation, a bill you could hold in your head. What it got wrong: backups were a prayer, updates were roulette, failover was one tired human at 2am.
Whereas the industry fixed the second list by surrendering the first. Give up ownership, receive operations.
Nobody tried the other direction.
Keep the ownership.
Fix the operations.
Now, therefore, this deed records what that direction looks like, built.
Part III The grant
Selfhost.dev grants the undersigned the following, in plain language, with the numbers attached.
Clause 1 of 6
Sight.
You are entitled to see the machine. A Project runs on a dedicated server in full view: live CPU, memory, and connections, streaming deploy logs, every service visible on one canvas. Not a black box. A glass box.
verified: dedicated single-tenant project servers · live metrics and full logs · deploy from GitHub with auto-deploy on push, PR previews, custom domains
Clause 2 of 6
Title.
You are entitled to keep what is yours. Managed databases run in our cloud or, with BYOC, inside your own AWS account under your keys. The engines are standard PostgreSQL, MySQL, Redis, and ClickHouse: your data never learns a proprietary dialect.
verified: BYOC in your own AWS account · PostgreSQL 16-18, MySQL 8.0-8.4, Redis 7.0-7.2, ClickHouse 24.3-25.1 · pgvector and TimescaleDB, one click on managed PostgreSQL
Clause 3 of 6
Consideration.
You pay for hours the machine actually runs. Credits are prepaid and visible, the meter is hourly, and at zero your resources pause, then restart when you top up. Nothing here bills you for the privilege of forgetting it.
verified: project servers $0.02-$0.11/hr (2 to 16 vCPU) · managed databases from ~$5/mo · top up from $5 · no tiers, no subscriptions
Clause 4 of 6
Agency.
Your infrastructure learned to read. 161 MCP tools let the editor you already work in provision, back up, scale, and diagnose all of it in natural language. The AI works where you work, and it answers to the owner. That would be you.
verified: 161 tools across 25 modules · Claude, Cursor, Windsurf, any MCP client · destructive operations require explicit confirmation
Exhibit C · Representative session · Tool names real
> what are we running right now?
list_instances
2x postgresql (multi-az) · 1x redis · 1x clickhouse · all healthy
> restore yesterday evening's state to a new instance
restore_pitr target 18:00:00 · new instance: api-db-restore
> drop the staging database
delete_instance
! destructive operation. type the instance name to confirm:
Clause 5 of 6
Exit.
Leaving must remain possible, or none of the above is ownership. Standard engines mean standard dumps. Moving in is white-glove and free: our engineers plan and run the migration with a tested restore before cutover. And there are no tiers to deprecate out from under you: your credits are yours.
verified: free migration help on every size · standard tooling in and out (logical replication, pg_dump) · no plans to deprecate
Clause 6 of 6
Exclusions, stated plainly.
We do not hold a SOC 2 report today, and we will not decorate this page with an invented uptime percentage. We would rather show you machinery: Multi-AZ failover on every managed engine, point-in-time recovery on the SQL engines, encryption in transit and at rest on managed instances, automated backups you can actually restore. When we earn the badges, they will appear here. Not before.
the full reliability story, mechanism by mechanism: /security/
Part IV The parties
A deed is only as good as the people whose names are on it.
M. Aziz
Infrastructure & databases
Years in startups on the systems side. Most recently led a DevOps team, driving cloud cost-optimization and building the CI/CD pipelines an engineering org shipped on. He knows how systems behave under real load and is genuinely obsessed with databases. The managed-database and no-surprise-cost core of Selfhost.dev is his to own.
Divyam Dhadwal
Product, AI & growth
8+ years building software, from enterprise mobility apps used by global brands and a Fortune 500 to running a startup's entire B2C engineering team. At Selfhost.dev he owns everything a customer touches, including the page you are reading and the MCP server your editor talks to.
Dhanraj Kshirsagar
Quality & backend
10+ years across startups and mid-size companies. Grew from QA engineer to senior QA, and now mentors QA teams while writing both the test automation and the backend code behind the platform. He is the reason reliability here is boring, in the best possible way.
Somya Hallan
Marketing & content
A marketing specialist with a doctorate who also runs her own content brand. She writes the Selfhost.dev blog and runs content and growth end to end, from SEO and topical strategy to the stories that explain what we build.
Selfhost.dev is built and run by three technical co-founders who cover the whole stack, infrastructure, product, and quality, plus the people who help it grow. It is a small, remote team, so there is no support tier to escalate through: when something matters, you are talking to the people who built it. The company, Swatantra Cloud Systems, is registered in Pune, India, and has been building Selfhost.dev since 2025.
On the incorporation papers the company is not called Selfhost.dev. It is called Swatantra Cloud Systems. Swatantra: sva, self; tantra, system. One's own system. We named the company after the feeling, then spent the years since building it.
Witnessed by 250+ developers who have signed up.
Part V Execution
This infrastructure belongs to
The signature is for you. This page transmits nothing.
$1 welcome credit at signup, no card needed. About 5 days of the smallest managed database, or 2 days of a project server.
Schedule A Questions, answered on the record (frequently asked questions)
A.1Is Selfhost.dev the same as self-hosting?
A.2What is Selfhost.dev?
A.3Who built Selfhost.dev and why?
A.4What can I run on Selfhost.dev?
A.5How does pricing work?
A.6What is BYOC?
A.7What are MCP tools?
A.8How is Selfhost.dev different from AWS RDS?
A.9How do I get started?
Schedule B The capabilities, itemized
Managed databases
Production PostgreSQL (16-18), MySQL (8.0-8.4), Redis (7.0-7.2), and ClickHouse (24.3-25.1) on AWS, with Multi-AZ and forking on all four, and PITR and pooling on the SQL engines. Migrate in with standard tooling, nothing proprietary.
Projects (a managed server we run)
Deploy from GitHub with auto-deploy on push, PR preview environments, custom domains, and live metrics.
Databases & templates
PostgreSQL (with pgvector for embeddings), Redis, MySQL, and MongoDB inside a project, plus a one-click Supabase template.
Bring your own cloud
Run your managed database in your own AWS account. Your data never leaves your infrastructure, no lock-in.
AI-native control
161 MCP tools across 25 modules. Manage everything from Claude, Cursor, or any MCP client.
Pay as you go
Prepaid credits, no tiers. Resources pause at a zero balance instead of generating a surprise bill.
Own your infrastructure.
Skip the operations.
A managed database from ~$5/month. A full project server from ~$0.02/hour. Pay only for what you run. See how we compare.