🧑‍💻 Solo & SMB 🤖 Native MCP ⚡ 2 min

Manganese PaaS

A complete app server in 2 minutes:
Dev · Test · Staging · Prod — no ops, no K8s, secure and sovereign

devtest stagingprod

Everything you need, nothing superfluous

PostgreSQL · HTTPS · Web IDE · AI MCP · Sunray Zero Trust · Sovereign by design

Ready in 2 to 10 min

Ubuntu LTS · Dedicated PostgreSQL · Automatic HTTPS · VS Code in-browser · systemd services.yml. Git repo cloned at provisioning. Zero ops configuration needed.

🔢

Dev · Test · Staging · Prod

Four distinct qualifiers to separate your environments mechanically. The prod qualifier activates approval gates automatically — no convention to follow, it is enforced in code.

🤖

Agentic MCP surface (~35 methods)

An AI agent drives all your environments: provisioning, deploy, exec, file I/O, journalctl, systemd control. Your assistant codes directly on the server, without a console.

🔒

Zero Trust Sunray included

Access to each environment protected by WebAuthn passkey (Sunray). Auto-generated config from Manganese. Secure access from any browser, no VPN, no IP to whitelist.

🔄

Reversible by design

Portable templates, native pg_dump, standard services.yml. Reproduce an identical environment on any host — or on your own infra. Zero platform lock-in.

🇫🇷

Documented sovereignty

Hosted in France on bare-metal (OVH, Hetzner…) or on your own server. Zero US cloud dependency in the critical path. 8 sovereign S3 backends.

🗄️

Dedicated PostgreSQL per molecule

Dedicated database in each environment, credentials injected as env vars (PGHOST, PGUSER, PGPASSWORD…). pg_dump and restore are first-class — staging → prod promotion in minutes.

📦

S3 multi-provider storage

AWS, MinIO, OVH, Scaleway, Cloudflare R2, Infomaniak, Wasabi, Dell ECS. Per-bucket credentials. Change provider without touching your application code.

Web
DB
S3
IDE

The Molecule concept PaaS

A Manganese molecule is a complete, autonomous application unit. It runs in an isolated LXC container on bare-metal, with its own compute resources, PostgreSQL database and S3 bucket.

Provisioned in 2 to 10 min from a Git repo. Each molecule contains:

  • Compute – Ubuntu LTS on bare-metal LXC, language-agnostic by design (Node, Python, Java, Go, PHP…)
  • Database – Dedicated PostgreSQL with credentials injected as env vars
  • Storage – S3 bucket on the provider of your choice (8 sovereign backends)
  • Integrated IDE – VS Code Server (Code Server) accessible from the browser
  • Automatic HTTPS – Traefik reverse proxy + included public TLS certificate
  • Git connection – Token validated and repo cloned at provisioning

Key advantages

What Manganese PaaS concretely changes

1

Zero ops to learn

PostgreSQL, HTTPS, systemd, reverse proxy — all pre-configured. Connect to your server and start coding. No devops required to get started.

2

Four environments, one platform

Dev, test, staging and prod live on the same platform, with the same template. What works in staging works in prod — same Ubuntu, same PostgreSQL, same env vars.

3

AI as ops co-pilot

Claude, Cursor, or any MCP client can provision, deploy, read logs and fix issues on your servers. Without opening a terminal, without writing a script.

4

Simple prod, not fragile prod

The prod qualifier activates approval gates on destructive operations. You stay in control without K8s infrastructure. For critical HA prod → ProdFleet HA.

5

Resize in 2 minutes

Need more RAM or CPU? mgx_resize_server — no migration, no extended downtime. The environment restarts on the new plan.

6

Coherent single-author stack

Manganese, Sunray and IMQ are open-source and developed by the same team. Cross-component edge cases are resolved internally — no fragile integration between third-party SaaS.

Use cases

Manganese PaaS in practice

🧑‍💻

Solo developer — full stack in 2 min

Create a dev molecule, connect your repo, open VS Code in-browser. All the tooling (Postgres, HTTPS, systemd) is there. No local docker-compose to maintain, no ngrok tunnel.

👥

Small team — dev + staging separated

Each developer has their own isolated dev environment. The staging environment is shared for validation. Promoting dev → staging → prod is done by changing the qualifier.

🏢

SMB — from demo to prod on the same platform

Start with XXS for the demo, scale to M for production. Same template, same platform, same deploy process. No migration between dev and prod environments.

🔁

Freelancer — dedicated client environment per project

Provision one molecule per client, with a staging qualifier for acceptance testing and prod for go-live. Deprovisioning at project end takes 30 seconds.

⚠️ Scope of this offer

No high availability — single LXC container per environment, single-node
No auto-scaling — manual resize available via mgx_resize
No cross-DC PostgreSQL replication — snapshots + pg_dump to S3 recommended
For critical production, HA and multi-DC → ProdFleet HA

Pricing

From the individual developer to the Enterprise with their own server

Manganese On Demand

0.09 €

per hour & per VM type

  • Multi-tenant (isolated VM)
  • PostgreSQL included
  • S3 included
  • HTTPS + Traefik
  • Integrated VS Code
  • AI MCP access
Get started

Manganese SaaS & Dedicated Public Server

0.2 € GB / month

of physical server RAM

  • Oursbl.eu leases the server *
  • Dedicated LXC (non-shared)
  • PostgreSQL HA included
  • S3 multi-provider included
  • Sunray Zero Trust included
  • Full MCP access

* Server: provider rate +20 %

Choose Dedicated

Manganese SaaS & Server On Premise

0.4 € GB / month

control plane only

  • You bring the internal server
  • Control plane only
  • Dedicated LXC (non-shared)
  • Netbird transport (EU solution)
  • Full MCP access

HA → switch to ProdFleet HA

Contact us

For HA and Enterprise offers → Manganese ProdFleet HA

Your first server in 2 minutes?

Solo developer · Small team · SMB · Mid-market · Sovereignty by design