Free Tools · Open source
OpsMaxx
Every server you look after, in one window.
- MCP tools for AI agents, each governed separately
- 28
- fleet operations across every server
- 20+
- database engines, through a bastion if needed
- 5
- free, no account, no telemetry, no paid tier
- MIT
Overview
OpsMaxx is a Swiss Army knife for sysadmins, infrastructure engineers and DevSecOps engineers. Terminal, SFTP, five database engines, tunnels and VPNs, an encrypted vault and fleet operations live in one desktop app, organised into workspaces per client or environment that share one credential store. Its MCP bridge lets AI agents such as Claude Code work on your servers by name, under allow, ask and deny rules, without ever seeing a credential. Free and open source under MIT, with no account, no telemetry and no paid tier.
How it works
From setup to enforcement
- Step 01
Download and open
Signed and notarised on macOS (Apple Silicon and Intel), plus Windows and Linux, or install with Homebrew. No sign-up and no licence key.
- Step 02
Bring your servers
OpsMaxx imports ~/.ssh/config, ProxyJump entries included, so the servers you already reach by name are there on first run.
- Step 03
Organise by workspace
One workspace per client or environment, each with its own servers, databases, tunnels and secrets, optionally password-locked and exported as one encrypted file.
- Step 04
Connect your agent
One command pairs Claude Code or Codex with the local MCP bridge. Give the agent a read-only group on staging and an ask-first group on production.
Features
Key features
Workspaces
Servers, databases, tunnels, vault entries and agent sessions scoped per client or environment, and an agent cannot see past its workspace.
Terminal and SFTP, one connection
GPU-rendered terminal with split panes and search, unlimited jump hosts, and an SFTP browser on the same session, so two-factor is typed once.
Five database engines
Query and shell into Postgres, MySQL, SQL Server, MongoDB and Redis, through a bastion when that is the only route.
Tunnels, VPN and inspection
Local and remote forwards, SOCKS5, userspace WireGuard with no admin rights, OpenVPN, frp, and a proxy that shows the HTTPS a machine really makes.
Encrypted vault
AES-256-GCM store for logins and API keys under a scrypt-derived master password. No MCP tool can read it.
Know your fleet
Inventory, configuration drift, capacity trends, security posture (SSH config, sudo rules, ports, firewall), key access and fleet-wide search.
Change safely
Patching in waves that stops at the first unhealthy server, fleet-wide commands, cron and timers, rules, verified backups and a change log. Writes are off until you turn them on.
Operate containers and databases
Docker, Compose drift, Kubernetes cordon, drain and exec, replication lag, slow queries, multi-server log tailing and runbooks.
MCP bridge for AI agents
Agents address servers by friendly name; every tool is ALLOW, ASK or DENY; output is redacted; escalation shells are always refused; every action is audited.
Who it's for
Built for how teams actually work
On call at 3am
An alert fires; the fleet monitor shows which server, the runbook shows what was run the last three times, and the terminal is one click away.
Consulting across clients
One password-protected workspace per client, each with its own servers and secrets, exported to a single encrypted file.
Keeping a platform patched
Patch in waves, watch drift since last week, and read security posture as it actually is on the box.
Working alongside AI agents
Let Claude Code work on staging on its own and wait for your approval on production, with every action in the audit log.
Technical details
Works with your stack
- Platforms
- macOS (Apple Silicon and Intel, signed and notarised), Windows, Linux; Homebrew cask
- Connectivity
- SSH with unlimited jump hosts, SFTP, SOCKS5, WireGuard, OpenVPN, frp
- Databases
- Postgres, MySQL, SQL Server, MongoDB, Redis
- Containers
- Docker, Docker Compose, Kubernetes
- Secrets
- Server credentials in the OS keychain; vault in AES-256-GCM with a scrypt-derived master password
- MCP clients
- Claude Code, Claude Desktop, Codex, Gemini CLI, any MCP client
- Agents never receive
- SSH passwords, private keys, database credentials, hostnames, IPs or usernames, vault contents, or an interactive root shell
- Licence
- MIT; no account, no telemetry, no paid tier
Pricing
Plans
Free and open source under MIT. No account, no telemetry, no paid tier.
Everything
Every feature, on every platform.
- No account or licence key
- No telemetry or analytics
- No session limits
- Full source on GitHub
FAQ
Questions teams ask
Is it really free?
Yes. MIT licensed, with no paid tier, no session limit and no subscription. The full source is public.
Does it phone home?
No account, no telemetry, no analytics. Every connection it makes is one you configured, apart from an update check you can switch off.
Can an AI agent do something I did not intend?
Only within the access group you set. Escalation shells are refused for every group, ASK actions wait for your approval, and every action lands in the audit log.
Where are my passwords and keys stored?
Server credentials go in your operating system keychain; the vault is AES-256-GCM under a master password that is never stored. Neither leaves your machine.
Products
30 min
Google Meet