ServerOps Search your VMs…
- prod-web-01linux
- prod-api-02linux
- db-primarylinux
- win-app-01windows
- gpu-node-01linux
>_
Self-hosted control plane
Deploy any software. Run, secure and understand it all.
One self-hosted control plane that takes any app from a Git repository to a live domain, then watches the servers, containers and databases under it — with security testing, error tracing, load and disk intelligence, project management, IT spend and an AI assistant built in.
130+modules in one control plane
Any stackPython to .NET, Docker, Windows
8 stepsfrom repository to live domain
Inside the control plane
Seven layers. One place to run them.
- 01 · Deploy
Repository to live domain.
Start a new project, import one from GitHub, or add a folder that already runs on a server. The wizard asks only what it needs and shows every step as it happens.
- 8 guided steps
- Resume a failed step
- 02 · Fleet
Every server. One agent.
Health, CPU, memory, disk, GPU, Docker, firewall and network per server, with alerts that explain themselves.
- Linux + Windows
- Browser terminal
- 03 · Data
Databases without the CLI.
MySQL, PostgreSQL, MongoDB, Redis, OpenSearch, Neo4j, Qdrant and SQLite — queries, index usage, users and config with automatic rollback.
- Read/write limits
- restic backups
- 04 · Security
Check before someone else does.
A hardening score with one-click fixes, exposure and TLS scans, and traffic capture of a service through an intercepting proxy.
- Hardening score
- TLS scans
- 05 · Insight
Know why, not just that.
Why a server is slow, where the disk went, which exception your users hit, and what the cloud accounts are running — explained from real numbers.
- Load analysis
- Disk intelligence
- Error tracing
- 06 · Projects
A project system, and the money.
Tasks, roadmap, docs, deploy history and per-project access — with every IT expense, recurring bill and budget tracked in the same place.
- Roadmap & tasks
- Expense ledger
- Roles & audit
- 07 · Assist
Your day, and an AI that helps.
Notes, reminders, time tracking, runbooks and a personal vault — plus an advisory AI assistant and plain-words search across everything in Studio.
- Productivity hub
- AI Copilot
- Ask Studio
Deploy
From a repository to a live domain, without a runbook.
Start a new project, import one from GitHub, or add a folder that already runs on a server. The wizard asks only what it needs and shows every step as it happens.
Deploy any type of software
- Flask
- FastAPI
- Django
- Node.js
- Next.js
- React · Vue · Vite
- Spring Boot
- Go
- Rust
- .NET
- Docker Compose
- Static sites
- Windows services
- 19 one-click apps
- Monorepos
- Your own commands
- 01
Name
Always stored in capitals so it reads the same everywhere.
- 02
Code & GitHub
Pick a repo from GitHub, or paste any git URL.
- 03
Server & folder
Browse the server, create folders open to every user.
- 04
App type & run
Flask, FastAPI, Django, Node, Next.js, Spring Boot, Go, Rust, .NET, Compose…
- 05
Secrets
Kept encrypted in Studio and delivered to the service, never in the repo.
- 06
Domain & SSL
Create the sub-domain at GoDaddy, the Nginx site and the certificate.
- 07
Auto-deploy
A push installs changed libraries, builds, migrates and restarts.
- 08
Launch
Review the plan, press once, watch it run — resume if a step needs a fix.
Lead with a domain
Everything linked, automatically.
A project, its service, its domain, its certificate and its push trigger — set up together, so the first thing you share is a working address.
- A deploy script that is more than
git pull— it reinstalls only when requirements.txt, pom.xml, package-lock.json or go.sum changed. - Guided CI/CD for existing projects — a checklist of what is in place, why it matters and where to look, then sets up what is missing.
- Failed deploy? The log shows the step that failed, not just an exit code.
- Secrets stay in Studio — view and edit them later; every reveal is recorded.
Operate from anywhere
Five consoles. Always one click away.
- 01 · Terminal anywhere
A terminal on every page.
A floating launcher sits in the corner of every Studio screen. Search your servers, open a real terminal or the file manager, and keep working where you were — tile windows, broadcast one command to many servers, forward ports.
- Floating launcher
- Broadcast
- SFTP files
- Port forwards
- 02 · Logs anywhere
Every log line, from anywhere.
Search every captured line across the whole fleet at once — by text, server, service and severity — or switch on live tail. Saved searches and shareable links put the answer one click away for the whole team.
- Fleet-wide search
- Live tail
- Saved searches
- Windows service logs
- 03 · AI insights
Logs that explain themselves.
The AI Copilot reads the same log index, alerts and metrics, and investigates on its own every hour. It names a likely cause, shows the lines it used as evidence and suggests a fix — advisory only, never invented numbers.
- Hourly investigations
- Evidence lines
- Ask in plain words
- 04 · Traffic
See who is hitting what.
Studio turns every service's Nginx access log into requests, error rates and response times — per service, per endpoint and per status code, across the fleet — with saved views and a drill-down for each service.
- Per endpoint
- Status codes
- Error-rate trends
- 05 · Security suite
Test apps and servers first.
TLS and header audits, port, exposure and CVE scans against the services and VMs you already registered, traffic capture through an intercepting proxy and opt-in load tests — every finding in one list, every scan in the history.
- Full audit
- CVE scans
- Intercepting proxy
- Load tests
Jugnu · talk to your servers
Just say it. Studio does it.
Press Alt + J and speak — English, Hindi or Hinglish. Jugnu opens the page, finds the folder on the server, fills in the project wizard, and asks before he changes anything. Pick a sentence and watch.
Alt Jfrom any page · a deploy, restart or reload always waits for your “haan”
Dock
Pin a window to the right, top right or bottom right — or move the dock to another edge.
Your screen, your layout
Pin anything. Nothing overlaps.
The Copilot, the terminal, logs and files each have a pin button. Pin a second window to the bottom and the first becomes the top half; pin too many and every column narrows evenly. The dock keeps its own band, so it never covers what you read.
- One set of rules for every window — right, top right, bottom right, or float.
- The dock moves to any edge — the page and the windows move with it.
- The terminal button stays where it is, however many windows you pin.
Intelligence
Why is it slow? Where did the money go?
Studio does not stop at charts. It reads the numbers it already collects and tells you what is going on, in plain words — advisory only, never invented figures.
Why is this server loaded?
CPU, memory, disk and network in one view, with the busiest processes ranked and explained — so a spike has a name next to it.
Where did the space go?
Disk Intelligence scans the largest folders and files, buckets them into real categories with real gigabytes, and an AI summary says what to clean first.
Track your whole IT expense
Servers, domains, licences and cloud bills in one ledger: recurring templates, card, UPI or net-banking details, up to three attachments, monthly budgets and an overview.
See the error your users saw
Application exceptions and traces from your own apps through Traceway, installed in one click, with the failing step shown next to the request that caused it.
GCP and Firebase in view
Every Google Cloud and Firebase project a service account can see, next to your cloud instances, object storage and infrastructure-drift checks.
Know before customers do
Heartbeats for sites and servers, alerts and on-call, maintenance windows, and public status pages that show only what you choose.
Hover a layer to pull it out · click to pin · drag to turn
Projects
A full project system, in layers.
Every project is more than a repository: its work, its money, its servers and its history live together, with the right people able to see the right layers.
- Tasks, roadmap and health — priorities, owners, due dates and a project health view.
- Deploys and rollback — GitHub auto-deploy, a deploy history with the step that failed, one-click rollback.
- Budget and expenses — what the project costs each month against what was planned.
- Files, docs and secrets — documents, an encrypted .env, and a timeline of who changed what.
Docker & security
Containers, scanned and under guard.
See every container on every server, run and update Compose stacks, and test what they expose — from the same place that deploys them.
- Docker management — containers, images, ports, logs and Compose files, with one-click self-hosted apps.
- Security testing toolkit — TLS and header audits, port and vulnerability scans, and opt-in load tests.
- Hardening score — a server-by-server score with one-click fixes and firewall rules.
- Traffic capture — see the requests a service makes and receives through an intercepting proxy.
Drag to orbit · click a container · click empty space to rescan
AI Hub · your models, your keys
One gateway. Every model you run.
Serve a model from your own GPU, add your Claude subscription and a fallback, and hand each of your projects a key with its own models, rate and daily budget. Flip the switch to see what the local gateway changes: the request stops travelling to Studio and back.
- vLLM on any server, with a tuning advisor that tells you whether it will fit and how fast it will be.
- Keys per project: allowed models, rate, simultaneous requests, daily budget — usage counted per project and model.
- Streaming passes straight through, with keep-alive so long answers survive Cloudflare.
Keys, not loginsper project, revocable, with a daily budget
Checked on the spotthe agent holds only hashes of the keys
Counted, not guessedusage flows back to Studio per project and model
Fresh from the workshop
Small things that remove the daily friction.
The big features are above. These are the details people notice after a week.
Ask, don’t beg.
A refused button becomes a request to an administrator — with your reason — and approval grants exactly what was refused.
I deploy this on Friday and need to restart it
Request itapproved ✓
The sheet fills itself.
Link a row to a project and its server, domain and certificate dates appear — grey until you type over them.
ProjectServerDomainSSL endsCitizen Portal10.0.0.4:8080services.example.gov28-12-2026Tourism Site10.0.0.7:3000tourism.example.in14-01-2027
One compose window.
File on the left, live output on the right — on the hub, on a server and inside a project.
services:
web:
image: nginx:1.27
ports: ["80:80"]
db:
image: mysql:8
web:
image: nginx:1.27
ports: ["80:80"]
db:
image: mysql:8
web Pulling
3f4ca61 Pull complete
8c9b2a1 Downloading 56MB
db Pulling
a1c3d92 Pull complete
Container app-web-1 Started
Container app-db-1 Started
web Pulling
3f4ca61 Pull complete
8c9b2a1 Downloading 56MB
db Pulling
a1c3d92 Pull complete
Container app-web-1 Started
Container app-db-1 Started
3f4ca61 Pull complete
8c9b2a1 Downloading 56MB
db Pulling
a1c3d92 Pull complete
Container app-web-1 Started
Container app-db-1 Started
web Pulling
3f4ca61 Pull complete
8c9b2a1 Downloading 56MB
db Pulling
a1c3d92 Pull complete
Container app-web-1 Started
Container app-db-1 Started
It waits for DNS — and says what it waits for.
The wizard asks public resolvers and the domain’s own name servers, then the certificate authority, and tells you where a record must live if it is missing.
Adding the A record at the registrar
Asking the domain’s own name servers
Waiting for public resolvers
Certificate issued · Nginx reloaded
It listens the way you speak.
English, Hindi or Hinglish — Studio starts with the language that worked last time and switches by itself when what it hears fits badly. No language menu.
“Projects kholo”
“प्रोजेक्ट खोलो”
“Open the terminal”
Type the name to delete it.
Removing a service asks for its exact name, and only admins can.
old-service
Delete service
Everything above is one list among a hundred and thirty modules.
Windows services listed in a second, Nginx files read through the agent, wildcard certificates by DNS, tunnels with a stable name, a schema diagram from real foreign keys, installable as an app, filters you can tick several of. The handbook walks through every one.
- Install as an app
- Speed mode
- Multi-select filters
- Windows services
- Wildcard certificates
- Tunnels
- Schema diagram
- Excel & PowerPoint viewers
- Chrome extension
- Python · Node · Go libraries
After the first deploy
Everything else you run, in the same place.
Everything below works from the browser, with roles and an audit trail behind it.
A helper you can talk to
Press Alt+J and say "projects kholo" or "naya project banana hai". Jugnu opens the page, finds the folder on the server, fills the wizard and asks before he changes anything. He also explains any error from the exact request that failed.
The project register
The sheet every office keeps — people, server, domain, SSL, AMC and documents — filled in by Studio for every project it knows, with pick-lists, date pickers and CSV in and out.
Ask for access, don't share a login
A refused button opens a request to an administrator; approval grants exactly what was refused. Every server gets a shared group so deploys by different people never collide.
Servers at a glance
Health, CPU, memory, disk, GPU, Docker, firewall and network per server, with alerts that explain themselves. Linux and Windows through one agent.
Work on the server from the browser
A real terminal with a command helper, broadcast to many servers at once, split windows, port forwards, an SFTP file manager and live read-only sharing.
Domains, SSL and Nginx
Every site with its certificate and expiry, an Nginx editor with validation, GoDaddy and Cloudflare DNS, tunnels, and renewal hooks that reload your services.
Databases without the CLI
MySQL, PostgreSQL, MongoDB, Redis, OpenSearch, Neo4j, Qdrant and SQLite — queries, index usage, users with read/write limits, and config files with automatic rollback.
Back up and rebuild
Per-server restic backups with retention, portable encrypted archives, cloud snapshots, and a rebuild wizard for another server.
Check before someone else does
Hardening score with one-click fixes, exposure, TLS and vulnerability scans, firewall rules, and traffic capture through an intercepting proxy.
Logs, uptime and tracing
Search logs across the fleet, watch uptime, follow Windows service logs live, and see application exceptions and traces with a one-click Traceway install.
Test your API like Postman
Choose a project, and Studio scans its routes and fills in the parameters — marked mandatory or optional — then sends real requests, imports Postman or OpenAPI, and smoke-tests after every deploy.
Tasks, roadmap and timeline
A full project management system: tasks, roadmap, documents, per-project access, a timeline of changes, and reports on how the team is doing.
Every expense, one ledger
Recurring bills, card, UPI and net-banking payments, invoices and receipts attached, monthly budgets and an overview of where the money goes.
Your day in one hub
Notes, reminders, time tracking, standup, runbooks, a watchlist, a personal vault, a calendar and a My Day page that gathers what is yours.
Ask, and it explains
An advisory Copilot that sees the page you are on (never passwords), Ask Studio search in plain words, and a playground for the AI models you connect.
GCP, Firebase and cloud servers
Google Cloud and Firebase projects, cloud instances, object storage and CDN, and infrastructure-drift checks in the same sidebar as your own servers.
Roles, access and audit
Viewer, operator and admin roles, per-server and per-project access, two-factor sign-in, shared credentials, and an audit log of who did what.
On-call, alerts and status
Incidents with on-call, alert rules and maintenance windows, and public status pages that show only the services you pick.
A toolbox beside the servers
A Tools Hub for everyday utilities, import of hosts from ssh config, PuTTY and SecureCRT, and one-click apps like Grafana, n8n and Gitea.
Nginx where you want it
Choose which server holds a web service's Nginx file and certificate: Studio's own, the server where the app runs (Linux or Windows), or a Bridge Proxy, where a public-address server forwards over the VPN to an app on a local server that has no public IP.
Your models, your subscription, one place
Run vLLM on any server with every setting explained, a model advisor that sizes the settings to your GPUs and shows the speed your hardware can really reach, a model store with resumable downloads, live monitoring and measured speed tests. The old stand-alone manager imports in one click.
One key per project, counted per project
Other projects call Studio with their own key and use Claude, your local models or any OpenAI-compatible provider through one address, in either the Anthropic or OpenAI format. Per-key models, effort, answer length, rate and token budgets; usage by project, model and day; no prompts stored.
Try every setting, copy the request
A form with every setting and a plain explanation of each, file attachments, streaming, and the exact request as curl, Python, the Anthropic or OpenAI SDK, or Node. Plus ready API docs for the people building on it.
Pinned windows, SSL, Docker
A log and a terminal side by side, each pinned to the right or minimized; SSL and Nginx changes that no longer fail on permissions; Docker deploys that show live progress; and Docker Hub search as you type.
Contracts, AMC and audits, tracked
Work order and timeline dates, AMC and security-audit validity with reminders that tighten as the date nears, an upgradation history, the standard project documents with a completeness check, and a letters timeline you can download as one indexed PDF.
Attach files, pick the model, keep your chats
Drag in photos, PDFs and code, choose the model and how hard it thinks, see your plan usage, and keep your own saved chats. It reads the page and the file you have open, and can answer through the AI subscription you already have.
Reach a server at 100%
Emergency tools to calm, freeze or kill the process that is eating a server — on Linux and Windows — plus an agent that keeps answering under load, and a deploy that can be stopped instead of hanging.
Pages that update themselves
Server overview, jobs, services and GPU pages receive live updates instead of polling, with a top-services view you can sort by CPU or memory and a per-core breakdown.
Your own mail server
Install mail for your domain on one of your Linux servers: mailboxes, DNS records, certificate, a readiness check and a test message, all from the browser.
Paste from Excel, or upload a sheet
Add many tasks or logins at once from a grid, a paste from Excel, or an uploaded file with a ready template. Office logins tagged official are shared with the whole team, read-only.
Reminded every day until it is done
Reminders with a date range, snooze that keeps your schedule, and task reminders that close themselves when the task is completed.
Crawl a whole site
Follow a site's own links to find broken links, duplicate titles and descriptions, thin or slow pages, and pages nothing links to.
How it compares
Most tools do one layer. Studio runs all of them.
Self-hosted PaaS tools deploy containers, desktop managers drive a VPS over SSH, Docker UIs manage containers, observability suites watch them and hosting panels run websites. Studio does those jobs — and keeps going into logs, traffic, security testing, AI insight, projects and spend.
| Capability | ServerOps Studioself-hosted control plane | Server Compassdesktop VPS manager | Openshipopen-source PaaS | OneUptimeobservability & incidents | Coolifyself-hosted PaaS | Dokployself-hosted PaaS | CapRoverself-hosted PaaS | Portainercontainer manager | Pleskhosting panel |
|---|
Built inPartial, add-on or per-app onlyNot a core feature
Based on each product's public documentation as of October 2026 and simplified to one mark per cell; features change quickly and paid tiers or extensions may add more. Product names belong to their owners — check their sites for the current picture. Server Compass is a desktop app that manages servers over SSH; Openship can also run as a hosted cloud; OneUptime monitors and responds to incidents but does not deploy or manage servers.
Screens
A look inside.
Real screens from the running application (addresses and host names masked).
Built to be trusted
Your servers, your data, your rules.
Studio runs on your own infrastructure. It is designed so that powerful actions are deliberate and traceable.
- Self-hosted — nothing leaves your network unless you connect a service you chose.
- Secrets encrypted at rest and revealed only on purpose, with the reveal written to the audit log.
- Read-only by default — changing, restarting or deleting needs the right role, and risky actions ask first.
- Two-factor sign-in and per-server access grants.
- Everything logged — who did what, when, and to which resource.
Early Access Program
Explore it. Host it yourself.
ServerOps Studio is still being shaped. Tell us who you are and what you want to build with it — accepted applicants are given access to the repository to explore the code and run Studio on their own servers.
- Repository access — read the source, run it, and tell us what is missing.
- Self-hosting — your servers, your data, your own instance.
- A say in the roadmap — your purpose and expectations decide what we build next.
This page is an introduction only. The Studio we run for our own operations is a separate, private deployment and is not reachable from here.
Wherever your servers run
Every server, every region —
connected to one Studio.
Apply for early access to get the repository, explore it and host it on your own servers.