Learn more

How Opsaido works, what it can do, and how it keeps your systems safe.

Smart enough to do the work. Secure enough to trust with it.

🧠

Smart

Understands plain English across every system you run. It investigates the cause, explains what it found, proposes the fix, and follows the conversation like a colleague would.

🛡️

Secure

Nothing changes without your approval. Dry run first, roll back after, with a full audit log, SSO and roles. Self-hosted, so your data stays on your network, and the AI can run fully local. No telemetry: the only thing it reports is your licence. Security posture: every Linux and Windows server is scored against built-in security baselines (CIS-aligned) and your own tech spec, with fixes that go through the same approval.

Everything in one place

One app replaces the dozens of consoles you log in to every day and every week: hypervisors, firewalls, Microsoft 365 and Azure, DNS, backups, storage and more.

Dashboard

Everything at a glance

What's going well, what needs attention and every system you manage. Drag cards, pick widths, and put it on a wall TV in kiosk mode.

Assistant

Ask in plain English

Questions answered from live data, and changes planned for you, with the exact command shown before anything runs.

No code needed

Commands built for you

No scripting, no syntax to remember. Say what you want and Opsaido writes the PowerShell, shell commands, API calls and config files, shows them to you, and runs them once you approve.

Security posture

Spot trouble before it spots you

Ask “check the firewall for anything suspicious” or “how secure are we?”. Opsaido reviews logs and settings across your systems, explains what matters and fixes it with your approval.

Security baselines

Every host measured against a standard

Linux and Windows servers checked against built-in CIS-aligned baselines and your own tech specs, Docker hosts and containers against Docker best practice, Kubernetes with kube-bench, and a full OpenSCAP scan on request. Agentless and read-only.

Vulnerabilities

Patch what attackers actually use

Known vulnerabilities on every host and in the container images they run, ranked by CISA's known-exploited list, the chance of exploitation (EPSS), severity and whether the service listens on the network.

Troubleshooting

Find the cause, fast

Describe the symptom. Opsaido checks DNS, ports, TLS, HTTP, services, logs and resources step by step, tells you where it breaks, and proposes the fix.

Dry run

Rehearse before you run

Say “dry run” and every step is listed with what it would send. It checks targets, spots conflicts and catches mistakes.

Rollback

Undo a whole job

Settings are captured before each change. Roll back one change or everything a request did, in one click.

Self-heal

Alerts that fix themselves

When an alert fires, Opsaido investigates with read-only checks, finds the cause and proposes the fix for your approval.

Storage & NAS

Your storage, looked after

Synology, TrueNAS, Proxmox and Windows storage in one place: capacity, disk health, snapshots, shares and replication, with warnings before a disk or volume becomes a problem.

Monitoring

Every host, VM and container

Live CPU, memory, disk and network for every host, VM and container. At a glance shows your whole estate with a health score, and history charts let you ask why something was slow.

Topology

See how it all connects

Hypervisors, VMs, switches and the services talking to each other, down to the switch port, plus a world map of traffic.

Certificates

Never miss an expiry

Request from Let's Encrypt or your own CA, renew automatically and push to Windows, Linux, routers and reverse proxies.

Credentials

One key, every server

Generate SSH keys, push them to hosts, switch connectors over automatically and rotate everywhere in one click.

Remote access

RDP and SSH in the browser

No VPN client. A floating toolbar, per-connection RDP options and sessions you can pop out into their own window.

Smart home

Your home, in the same chat

Lights, heating, locks and automations through Home Assistant, in the same conversation as your servers.

Voice

Just say it

Talk to Opsaido in the browser, or through Home Assistant voice satellites and phones. It answers out loud, and changes can be approved with a spoken PIN.

Quick actions

Your common jobs, one click away

Turn any request into a button: a morning health check, restart unhealthy containers, patch servers as a dry run.

Workflows

Run it on a schedule

Any request can run daily, weekly or on a cron: a Monday patch report, a nightly backup check, a certificate expiry sweep. Results come to you.

Templates

Guided builds from templates

Step-by-step builds with sensible defaults: a new Active Directory domain, VMs from a template, a Docker app stack or a private AI server.

Teams

Ready for business

Roles and permissions, SSO or Active Directory / LDAP sign-in with group mapping and auto-created users, second approvals and a full audit log.

Host details

Every check, on every host

Open any host to see each check's required and current value. Fix it with approval or accept it with a reason, and follow the score over the last 15 scans.

PDF reports

Reports ready for the auditor

A clean PDF for every host and every scan: the score, failures by severity, results by section and exactly what needs fixing. Excel reports cover the whole fleet.

Software inventory

Every app, on every host

Every app and package on your Linux and Windows servers, and every container image your Docker hosts run, in one list: how many hosts have each one, which versions, and what needs attention.

Which host has what

See where every version runs

Open any app to see each host that has it and the version there. Hosts on an older version than the rest of the fleet, and versions with known vulnerabilities, stand out.

Container images

Every image, and where it runs

All the container images your Docker hosts run, with their tags, the host and container each one runs on, and the known vulnerabilities in them.

Virtualization

Your whole virtual estate

Proxmox, VMware, Hyper-V, XCP-ng and KVM in one place. Create, clone, migrate and snapshot VMs with guided wizards, manage datastores and networks, and keep clusters balanced with affinity rules and maintenance mode.

Import & export

Bring VMs in, take them anywhere

Import OVA appliances and disk images (qcow2, vmdk, vhd / vhdx, vdi, raw) from a link or straight from your computer. Export any VM for another platform, or move whole VMs from VMware, Hyper-V, Nutanix and XCP-ng with warm migration.

Backup & recovery

Back up, restore, prove it

Back up or restore any VM in a few clicks, manage Proxmox Backup Server jobs and Veeam, test a whole site recovery plan, and send a backup report that shows what's protected and what isn't.

Patching

Updates without the late night

A patching wizard for Linux and Windows servers, Proxmox and VMware hosts, Hyper-V, Kubernetes nodes and Docker hosts. It moves VMs away first, patches, reboots, checks health and puts everything back, one host at a time.

Documentator

Documentation that writes itself

A polished PDF for any server, VM, firewall, DNS service or your whole infrastructure: roles and configuration, Active Directory, certificates, Docker, backups, switch ports and the physical network, explained in plain words.

Network

Every switch port, live

See which machine is on which switch port, at what speed, on which VLANs, with live traffic. Works with any switch that speaks SNMP, and with Omada and UniFi, with alerts for ports that go down, slow links, errors and spanning tree changes.

Firewalls

Every firewall, one view

OPNsense, pfSense, FortiGate, Palo Alto, Check Point, Meraki, SonicWall, Sophos and Cisco side by side: rules, VPNs and certificates, and a live log of what's blocked and allowed across all of them.

Docker stacks

Compose without the terminal

Edit a stack's compose file and .env in the app, validate them, and save back to where the stack lives: files on the host, Portainer or Dockhand. Then deploy, pull from or push to its Git repository, or start a new stack from a repository.

IPAM

Know every address

See your whole address space: which IPs are used, free or reserved, which device had each one and when, with a history per address. Get an alert the moment an unknown device shows up on the network.

Recoverability

Prove your backups restore

Don't hope a backup works, test it. Opsaido restores to an isolated place on a schedule, checks the result, and scores how recoverable each system really is, with a report that shows what is protected and what is not.

Database backups

Back up databases and prove the restore

PostgreSQL, MySQL / MariaDB and SQL Server backed up on a schedule or by hand to a folder, an SMB share, NFS or S3. Every backup is verified, restored as a new database so the original is never touched, and proven by a test restore that counts toward your Recoverability score.

And plenty more

The smaller things that make every day easier.

Ask about anything

One assistant for every part of your stack. Pick an area to see a real conversation.

Ask Opsaido

1000+ things it can do, in 65+ areas

Every action the assistant can take, area by area. Mix them freely: one plain request such as “patch web-01, check last night's backup and post the result to Teams” can reach across several integrations, and the assistant plans the steps and runs them in order. Look-ups run straight away; changes wait for your approval and can be dry-run first.

How the assistant works

  1. 1

    Connect

    Add your systems in minutes: Proxmox, Windows, Active Directory, DNS, firewalls, NAS, Docker, Kubernetes, cloud providers (Microsoft 365, Azure, AWS, Hetzner, DigitalOcean) and more. Restricted networks connect through a gateway.

  2. 2

    Ask

    No code, no scripts. “Create an A record for app1.” “Why does docker-1 spike at 3 a.m.?” “Patch the Linux servers.” Plain English, from the browser, Telegram, Teams or Slack.

  3. 3

    Approve

    Ask for a dry run: every step is rehearsed and listed with the exact commands, and nothing changes. Happy with it? Approve the plan once, not every step; any variance asks again.

  4. 4

    Done

    It carries out the approved plan and checks each step. If a step varies from the plan or fails, it stops and asks you first. Every change is logged and can be rolled back.

Built for systems you can't afford to break

Powerful automation with the guard rails an IT team expects.

✓

Approval for every change

Reads run freely; anything that changes a system waits for you, showing the exact SSH command, PowerShell or API call.

✓

Works with or without the AI

Everything important has its own screen: virtualization, Docker, Kubernetes, databases, backup and restore, monitoring and alerts, security scans, patching, certificates, remote desktop and SSH, changes and rollback. Use the assistant for the hard parts, or run Opsaido with no AI provider at all.

🧪

Dry run

Say “dry run” and every step is rehearsed and listed. It checks targets exist, spots conflicts and catches mistakes before they happen.

📋

Nothing runs off-plan

Approve the dry-run plan once and it runs step by step, exactly as approved. Any variance, such as a different target, an extra step or a failed step, stops the run. It explains what changed and what it recommends, and nothing continues without your approval.

↺

One-click rollback

Settings are captured before each change, so a whole job can be undone at once. (Deletes are protected by backups instead.)

👥

Second approvals

Require another person for risky actions, per role, group or user. Approve from the web or straight from Telegram.

🔐

Roles, SSO and audit

Fine-grained permissions, Entra ID / Google / Okta / Authentik sign-in with group mapping, and a full audit log.

🔎

Security posture checks

Ask it to review the firewall, failed logins, exposed ports, admin accounts without MFA and missing patches. It explains what matters and fixes it with your approval.

🛡

Secrets stay secret

Credentials are encrypted, masked in every answer and log, and never sent to the AI. Run the AI locally if you prefer.

📵

No telemetry

We don't collect usage data, host names, logs, chats or results. The only call home is licensing: activation and a periodic check-in with your licence ID, installation ID, version and the number of machines. Offline licences never call home.

🗄

Your database, your choice

SQLite out of the box, or PostgreSQL for bigger installs. Move between them from Settings: every row is copied and counted before you switch.

✓

Proof your servers meet your standard

Upload your tech spec as Excel, Word or PDF and Opsaido checks every Linux and Windows server against it, alongside built-in CIS-aligned baselines. Read-only and agentless: nothing to install. Download the compliance report for your auditors.

Your AI, your choice

Choose from 25 AI providers, in the cloud or on your own hardware so nothing leaves your network. Opsaido can even build a private AI for you: chat, voice and a wake word on one GPU or CPU server.

  • Cloud: Anthropic Claude, OpenAI, Azure OpenAI, Google Gemini, Amazon Bedrock, Mistral, xAI Grok, DeepSeek, Groq, Cerebras, OpenRouter, Together, Fireworks, DeepInfra, NVIDIA NIM, Hugging Face, Moonshot Kimi and Alibaba Qwen
  • On your own hardware: Ollama, LM Studio, vLLM, LocalAI, llama.cpp, Jan or any OpenAI-compatible server. Nothing leaves your network
  • Pick the model per conversation, with usage reports and monthly token or spend budgets (warn or block)
  • House rules in plain English, and choose which tools it may use
  • Answers in your language: automatically in the language you write in, or always in one language for everyone
ClaudeGPTGeminiAzureBedrockMistralGrokDeepSeekOpenRouterOllamaLM StudiovLLMWhisperKokoro

Built to keep AI costs low

Most of an AI bill is what gets sent with every message, not the answers. Opsaido is set up to send as little as possible, out of the box.

  • Smart chat memory: long jobs send only what's needed each turn, recent steps and the approved plan in full and older output shortened, instead of the whole history every time
  • Cost-saving defaults on every provider you add: compact chat memory, low reasoning effort and a cap on answer length
  • Your controls per provider: chat memory, max tokens per answer, reasoning effort, temperature and Top P
  • No approval round trips for lookups: read-only checks run straight away, and an approved plan runs without asking at each step
  • Token meter in every chat: see what a conversation has used and cost
  • Mix and match: a cheap fast model for day-to-day questions, a stronger one for big builds, or free local AI
Smart chat memoryLow reasoning effortAnswer capMonthly budgetsUsage reportsFree local AI

No telemetry. Ever.

Opsaido doesn't report on you. What it does on your network stays on your network.

  • Never collected: usage data, host names, IP addresses, logs, chats, commands, results or credentials
  • The only call home is licensing: activation (licence key, installation ID, version) and a periodic check-in (licence ID, installation ID, version and the number of machines)
  • Offline licences never call home, for air-gapped and regulated sites
  • Your AI, your call: questions go only to the AI provider you choose, or nowhere at all with local AI
  • No analytics or trackers inside the app
No telemetryNo trackersCounts onlyOffline licencesLocal AI

Questions