60 FREE CONNECTORS · OPEN SOURCE · RUNS ON YOUR MACHINE
Your MSP tools, answerable by AI - without your data leaving your network.
Ask one plain-English question across every client - “which clients had backup failures last quarter?” - and get the answer back in seconds, not an afternoon of exports and pivot tables. Every connector is free, open source, and runs on your own machine: your tools stay where they are, and the answers come to you.
- Cross-client answers in one ask. One question spans your whole book of business - not one client at a time.
- Nothing to trust but your own machine. No cloud middleman, no servers of ours, credentials stay local. Your AI sees answers, not your raw data.
- Every command is safety-tiered - read, write, or destructive - so you control what an agent may do. See the Trust Center →
Find your tool → Already have an official MCP server? → View on GitHub →
A stranger filed a real OAuth bug at 9:14am. The fix shipped at 9:52am. That's the maintenance bar.
The overnight factory keeps every connector current - a hand-fix ledger checked in CI guarantees live-API fixes survive every regeneration.
How fixes outlive a reprint →Instead of exporting three reports and pivoting them in Excel...
just ask: “Which clients had backup-failure tickets last quarter?”
halopsa-cli tickets search --tag backup-failure --since last-quarterYour AI agent runs this for you - you never type it.
🛠 Built live with MSPs. Join a free weekly Build Session at compoundingteams.com/build-sessions - watch a connector built against a real MSP system, or bring your own.
One thing, many names. What this site calls an MCP server is what ChatGPT calls an app or connector, Claude on the web calls a connector, Microsoft Copilot calls a connector, and Claude Code calls a Skill. Same standard underneath: the Model Context Protocol.
My vendor already ships an MCP server. Why add this one?
Keep it - it’s good at what it does. A vendor’s own MCP server is excellent for working inside that one product: pulling up a record, drafting a reply, summarizing a single ticket. This is not a replacement for that.
An MSP Skills connector is the layer your vendors structurally can’t ship: the cross-client, cross-system layer. It keeps a private, searchable copy of your tool’s data on your own machine, so a question that spans every client and joins your PSA to your backup tool comes back as one instant local lookup - not hundreds of rate-limited API calls.
| A typical MCP server | An MSP Skills connector | |
|---|---|---|
| Where answers come from | A live API call per question | A private local copy of your data, plus the live API |
| Cross-client questions | One client at a time, paginated and rate-limited | The whole book of business in one ask |
| What the AI reads | Raw records streamed into the chat | Query results - the answer, not the bulk data |
| Systems per question | One (the vendor’s own product) | Many - join your PSA, backup, RMM, and M365 |
| Where it runs | Often a hosted cloud service | Your own machine - nothing leaves your network |
It is complementary, not a replacement: keep your vendor’s MCP server for in-product work, add an MSP Skills connector for everything that has to span your whole MSP.
Why MSP owners use this
Answers your vendor's API can't
A private, searchable copy of your data lives on your machine, so a 90-day, 10-client question is one instant local lookup - not hundreds of rate-limited API calls.
One quality bar, every connector
Each one passes mechanical verification before it ships. The badge tells you whether a real MSP has confirmed it against a live tenant yet.
Free, local, yours
Open source, Apache-2.0, runs on your own hardware. No per-tech fee, no SaaS lock-in, nothing to rip out.
Works with your agent
The six agents MSP owners actually use (self-serve, works today):
| Your AI agent | Why MSPs use it | Install |
|---|---|---|
| Claude Desktop | The most common MSP-owner choice - no terminal, just a chat window | Setup → |
| ChatGPT (Plus/Pro+) | The brand most MSPs already pay for; pair with Developer Mode | Setup → |
| Codex CLI (OpenAI) | Same audience as Claude Code, OpenAI side | Setup → |
| Claude Code (CLI) | For technical-leaning MSP owners or senior techs | Setup → |
| Claude Cowork | Anthropic’s desktop agent - paste a prompt, it installs itself | Setup → |
| GitHub Copilot (VS Code) | The Microsoft surface that runs the local binary today, in Agent mode | Setup → |
Microsoft & Google paths - big install base, but the honest remote / enterprise route:
| Your AI agent | What it takes | Install |
|---|---|---|
| Microsoft 365 Copilot / Copilot Studio | Remote only: host over HTTPS + Copilot Studio license + tenant admin | Setup → |
| Google Gemini | Gemini CLI: local. Gemini app: remote (HTTPS). | Setup → |
Skill-native agents (secondary): Hermes → and OpenClaw → read the skill directly and speak MCP. Also works with Cursor, Windsurf, Cline, Continue.dev, and Zed. Full per-tool deep-dive in which agent →.
What’s in the box
| Connector | What it does | Status |
|---|---|---|
| Abnormal Security → | Triage Abnormal email threats, confirm remediations, and pull client-ready security reports from your terminal. | Awaiting MSP receipt |
| Acronis Cyber Protect Cloud → | Spot failed backups, offline agents, and unbilled usage across every Acronis tenant by asking. | Awaiting MSP receipt |
| Action1 → | Patch every client’s endpoints and triage CVEs fleet-wide from your terminal. | Awaiting MSP receipt |
| Afi → | Answer fleet-wide Afi backup coverage, staleness, and licensing questions from one local store. | Awaiting MSP receipt |
| AppDirect → | Reconcile AppDirect marketplace billing and chase failed payments across every reseller company before month-close. | Awaiting MSP receipt |
| Atera → | Answer fleet-health, SLA, and book-of-business questions across your Atera estate from the terminal. | Awaiting MSP receipt |
| Autotask PSA → | Triage tickets, surface unbilled time, and project retainer burn by asking. | Awaiting MSP receipt |
| AWS → | Get a non-expert from zero to a Slack-delivered, plain-English, waste-flagged AWS bill breakdown - and cache it so you don’t pay per Cost Explorer call. | Awaiting MSP receipt |
| Axcient x360Recover → | Catch failed and stale Axcient backups, RPO breaches, and unbilled usage across every client by asking. | Live-verified |
| Better Stack → | Ask what’s down, who’s paged, where coverage is missing, and your MTTA/MTTR across your Better Stack account | Awaiting MSP receipt |
| Blumira → | Triage every Blumira finding across all your MSP client accounts from one ranked queue | Awaiting MSP receipt |
| CIPP → | Every client tenant’s M365 posture, license waste, and stale accounts in one cross-tenant rollup. | Awaiting MSP receipt |
| ConnectWise Automate → | Sync your entire ConnectWise Automate fleet into local SQLite and answer cross-client questions the per-server web UI can’t: fleet-wide health roll-ups, stale-agent sweeps, patch-compliance-by-client, and overnight drift. | Awaiting MSP receipt |
| ConnectWise Control → | Manage ConnectWise Control (ScreenConnect) remote support and access sessions from the terminal or your AI agent: list and inspect sessions across groups, run commands on guest machines, rename and tag sessions, manage instance users, and query the audit log - the whole instance from one CLI. | Awaiting MSP receipt |
| ConnectWise PSA (Manage) → | Triage boards, catch unbilled time, and prep QBRs by asking. | Awaiting MSP receipt |
| Cove Data Protection → | Sweep every Cove customer for failed and stale backups, storage growth, and month-end billing from your terminal. | Live-verified |
| CrowdStrike Falcon → | Triage Falcon alerts, find stale sensors, and rank critical vulnerabilities across every tenant by asking. | Awaiting MSP receipt |
| Datto BCDR → | Fleet-wide Datto BCDR answers - failed screenshots, stale backups, at-risk clients - from a local mirror in one ask. | Awaiting MSP receipt |
| Datto RMM → | Fleet-wide Datto RMM answers - stale agents, alert storms, patch and AV gaps - from a local mirror in one ask. | Awaiting MSP receipt |
| Domotz → | Roll up Domotz fleet health, offline devices, and asset inventory across every monitored site from the terminal. | Awaiting MSP receipt |
| Gradient MSP → | Push usage counts, catch billing drift, and trace alerts to PSA tickets in Gradient MSP Synthesize from your terminal. | Awaiting MSP receipt |
| HaloPSA → | Triage tickets, pre-empt SLA breaches, run cross-client analytics by asking. | Live-verified |
| HubSpot → | Pipeline health, stale deals, and owner workload for every client by asking. | Live-verified |
| Hudu → | Audit Hudu documentation hygiene, find stale passwords and expiring certs, and search every company offline. | Awaiting MSP receipt |
| Huntress → | Triage Huntress incidents and coverage gaps across every client org from one queue. | Live-verified |
| IT Glue → | Find any device, audit stale passwords, and rank thin client docs by asking. | Awaiting MSP receipt |
| Kaseya BMS → | Run the Kaseya BMS service desk, contracts, and billing from your terminal - with offline analytics the web grid can’t compute. | Awaiting MSP receipt |
| KnowBe4 → | Hunt repeat clickers, track risk drift, and close training-coverage gaps across every KnowBe4 client from your terminal. | Awaiting MSP receipt |
| Level → | Rank your riskiest endpoints, patch posture, and stale devices across your Level RMM fleet from one command. | Awaiting MSP receipt |
| Liongard → | See every config change, stale collector, and offline agent across your whole Liongard estate from one command. | Awaiting MSP receipt |
| Maxio → | Open, local revenue-intelligence CLI for Maxio Advanced Billing - MRR waterfalls, retention, and per-client history computed offline from a SQLite mirror, so the trended history survives even though the live API returns only point-in-time figures. | Awaiting MSP receipt |
| Microsoft Graph → | Find wasted M365 licenses, audit global admins, and triage Defender alerts by asking. | Awaiting MSP receipt |
| MSPbots → | Pull datasets, snapshot KPIs, and track ticket backlog trends by asking. | Awaiting MSP receipt |
| N-able N-central → | Find any device, triage issues, and search across every client by asking. | Awaiting MSP receipt |
| Nerdio Manager → | Audit autoscale, sweep idle session hosts, and reconcile per-customer billing across every Nerdio Manager account from your terminal. | Awaiting MSP receipt |
| NinjaOne → | Fleet-wide NinjaOne answers - patch compliance, backup gaps, AV blast-radius - from a local mirror in one query. | Live-verified |
| PagerDuty → | Triage PagerDuty incidents, on-call schedules, and escalation policies from the terminal. | Awaiting MSP receipt |
| PandaDoc → | Find stalled proposals, aging quotes, and unsigned PandaDoc documents from the terminal. | Awaiting MSP receipt |
| Pax8 → | Reconcile Pax8 billing, track MRR, and catch usage overages before they hit the customer invoice | Awaiting MSP receipt |
| Pipedrive → | Surface stale deals, weighted forecasts, and overdue follow-ups across your pipeline by asking. | Awaiting MSP receipt |
| Proofpoint TAP → | Investigate Proofpoint TAP threats, VAPs, and top clickers from your terminal without burning the daily API quota. | Awaiting MSP receipt |
| QuickBooks Online → | Answer AR/AP aging, DSO, cash-forecast, and month-end close questions across QuickBooks Online by asking. | Live-verified |
| Resource Guru → | Every Resource Guru endpoint as a typed command, plus a local database and per-resource per-day utilization no Resource Guru report gives you. | Awaiting MSP receipt |
| Rewst → | Operate and monitor Rewst RPA across every client org from the terminal or your AI agent: workflow execution health, failure triage, automation ROI, dormant-workflow cleanup, cross-org config drift, and integration-pack coverage - the multi-tenant answers the web console makes you assemble one org at a time. | Awaiting MSP receipt |
| RocketCyber → | Triage your managed SOC, rank devices at risk, and audit suppression rules from the terminal - no console logins. | Awaiting MSP receipt |
| Rootly → | Every Rootly incident, alert, and on-call object as a typed command, plus offline MTTR and coverage-gap answers. | Awaiting MSP receipt |
| runZero → | Find every asset on your attack surface, diff it over time, and trace any CVE to the machines it hits. | Awaiting MSP receipt |
| Salesbuildr → | Surface stale quotes, thin-margin lines, and pipeline velocity across your Salesbuildr quoting by asking. | Awaiting MSP receipt |
| SentinelOne → | Triage threats, find protection gaps, and catch overnight fleet drift across every SentinelOne tenant by asking. | Live-verified |
| Servosity → | Fleet attention, stale backups, QBR reports, and restore watch for every client by asking. | Live-verified |
| Sherweb → | Reconcile what you owe Sherweb against what customers owe you and surface true net margin per customer. | Awaiting MSP receipt |
| SkyKick → | Catch silently-failing Microsoft 365 backups across every SkyKick tenant before a customer needs a restore. | Awaiting MSP receipt |
| SuperOps → | Catch SLA breaches, unbilled time, and at-risk assets across PSA and RMM by asking. | Live-verified |
| Syncro → | Run every Syncro PSA and RMM workflow from your terminal: tickets, billing, and patch gaps in one ask. | Awaiting MSP receipt |
| Tactical RMM → | Triage every agent, sweep patch posture, and run a script across a filtered cohort by asking. | Awaiting MSP receipt |
| ThreatLocker → | Clear approval backlogs, beat the 31-day audit cliff, and find dark agents across every tenant by asking. | Awaiting MSP receipt |
| Veeam → | Every Veeam Service Provider Console endpoint as a typed command, plus a local multi-tenant mirror, cross-company rollups, and drift detection no per-instance Veeam tool offers. | Awaiting MSP receipt |
| WordPress → | Publish and manage WordPress pages, posts, media | Awaiting MSP receipt |
| Xero → | Answer AR/AP aging, reconciliation, and GL tie-out for a Xero org from the terminal - offline, in one call. | Awaiting MSP receipt |
| Zammad → | Read, search, and triage Zammad tickets and Knowledge Base, plus team-management analytics the API can’t answer in one call. | Awaiting MSP receipt |
New to the term? What is an MCP server? →
Let your AI pick for you
Already have msp-skills installed? Let the concierge recommend the right connectors for your stack and install the ones you approve. Paste this into Claude Code or Codex:
You have msp-skills installed. Using everything you know about me and how I work, recommend which connectors I should install - and install the ones I approve.
More in the concierge README →.
What makes this different
Local mirror, not live calls
Most MCP servers for PSA and backup systems proxy each question into a live API call. That’s fine for one ticket. It dies at QBR time, when you’re asking “how many backup-failure tickets across all 47 clients last quarter?” - that’s 47 paginated REST calls, rate-limit headaches, and a context window full of raw JSON the model has to re-read. These connectors sync your tool into a local SQLite mirror with full-text search. Cross-client and cross-engine questions become one local query: instant, offline, and the AI sees the answer, not the raw data.
Works with the AI you already use
You don’t have to switch agents. Same package, two interfaces: a Skill for Claude Code / Codex, and an MCP server for Claude Desktop, ChatGPT, Cursor, Windsurf, Cline, Continue.dev, Gemini, and Copilot. One install drops both binaries on your machine. Use one or both - your call. No vendor lock-in, no proprietary plugin format, no SaaS subscription that ties you to a single AI.
MSP owner first, not developer first
You don’t have to know JSON, regex, or what “stdio transport” means. Paste one sentence into Claude Code or Codex and your agent reads the Skill, installs the binary, and walks you through authentication. Every command has a tier (read / write / destructive) and a recommended agent policy in each skill’s governance.md - reads are always safe, and the recommended agent rule is to preview writes with --dry-run and require your approval before any mutation. The hard stuff is hidden; the safety bar is high.
Install in 60 seconds
Path A - let your AI install it (recommended)
Not sure which connector to pick? Let your agent decide. Paste this into Claude Code, Codex CLI, or Claude Cowork:
Read https://github.com/Servosity/msp-skills and, using everything you know about me and how my MSP works, recommend which connectors I should install - then install the ones I approve.
Or install a specific connector. The Servosity prompt:
Install the Servosity Skill and MCP server from Servosity/msp-skills in this agent workspace. If this workspace uses a POSIX shell (macOS, Linux, WSL, or Bash), run
bash <(curl -fsSL https://raw.githubusercontent.com/Servosity/msp-skills/main/skills/servosity/install.sh). If it uses Windows PowerShell, runiwr -useb https://raw.githubusercontent.com/Servosity/msp-skills/main/skills/servosity/install.ps1 | iex. Then authenticate withSERVOSITY_MSP_TOKEN=<your-partner-token> servosity-cli doctorand runservosity-cli --helpto explore.
And the ConnectWise PSA (Manage) prompt:
Install the ConnectWise PSA (Manage) Skill and MCP server from Servosity/msp-skills in this agent workspace. If this workspace uses a POSIX shell (macOS, Linux, WSL, or Bash), run
bash <(curl -fsSL https://raw.githubusercontent.com/Servosity/msp-skills/main/skills/connectwise-manage/install.sh). If it uses Windows PowerShell, runiwr -useb https://raw.githubusercontent.com/Servosity/msp-skills/main/skills/connectwise-manage/install.ps1 | iex. Then authenticate per the README and runconnectwise-manage-cli --helpto explore.
The same prompt works for every connector in the table above - swap the skill name and slug. If your agent can’t run shell, use Path B below.
Path B - run the installer yourself
Servosity on macOS / Linux:
bash <(curl -fsSL https://raw.githubusercontent.com/servosity/msp-skills/main/skills/servosity/install.sh)
Servosity on Windows (PowerShell):
iwr -useb https://raw.githubusercontent.com/servosity/msp-skills/main/skills/servosity/install.ps1 | iex
Swap servosity for any connector’s slug from the table above (abnormal, acronis, action1, afi, appdirect, atera, autotask, aws-billing, axcient, betterstack, blumira, cipp, connectwise-automate, connectwise-control, connectwise-manage, cove, crowdstrike, datto-bcdr, datto-rmm, domotz, gradient, halopsa, hubspot, hudu, huntress, itglue, kaseya-bms, knowbe4, levelio, liongard, maxio, microsoft-graph, mspbots, n-central, nerdio, ninjaone, pagerduty, pandadoc, pax8, pipedrive, proofpoint, quickbooks, resourceguru, rewst, rocketcyber, rootly, runzero, salesbuildr, sentinelone, servosity, sherweb, skykick, superops, syncro, tactical-rmm, threatlocker, veeam, wordpress, xero, zammad). New to all this? See the first-time setup guide →.
Now what? Once your connector’s
--version(ordoctor) returns clean, bring your tenant + your hardest cross-client question to a free Build Session - we’ll work it live with the MSP cohort and the same connectors you just installed.
Frequently asked questions
Does this work with ChatGPT?
Yes, on Plus, Pro, Team, Business, Enterprise, and Education plans (Free tier does not yet expose Developer Mode). ChatGPT connects to remote MCP servers over HTTPS, not local binaries directly. These connectors ship local binaries, so to use them with ChatGPT you run them on your machine and expose them via the mcp-remote bridge or your own HTTPS endpoint. Step-by-step: ChatGPT integration →.
Does this work with Claude?
Yes - Claude Code, Claude Desktop, and the Claude.ai web (via Claude Desktop’s MCP). Claude Code reads the Skill directly; Claude Desktop loads each connector as an MCP server you add to claude_desktop_config.json. Both paths are first-class. Most MSP owners start with Claude Desktop because it’s a Mac/Windows app - no terminal.
Do I need to know how to code?
No. The recommended install path is to paste one sentence into Claude Code or Codex - your agent reads SKILL.md and does the install. The fallback is a one-line installer per OS (bash or PowerShell). Neither path requires writing code, editing JSON in a text editor, or installing Node, Python, Docker, or a database. You will need to enter API credentials your PSA or backup tool gives you.
Is my PSA, CRM, or backup data safe? Does it leave my network?
Your data stays on your machine. Each connector runs locally: the CLI and the MCP server are binaries on your laptop. The SQLite mirror sits in a local directory under your user account. The AI agent (Claude, ChatGPT, Codex) only sees what the CLI or MCP server returns - typically the result of a query, not raw bulk data. Credentials are read from your environment or your agent’s config; they’re never bundled into the repo or transmitted to our servers (there are none).
Will this hit my vendor’s API rate limits?
Almost never. Each connector syncs your data into a local SQLite mirror once, then incrementally; subsequent questions (triage, SLA breaches, client cards, cross-client analytics) run against the local mirror, not the live API. The big-batch QBR queries that get you 429’d with API-passthrough tools become a single SQL join here.
How is this different from my vendor’s built-in AI?
Vendor-native AI is great for ticket-by-ticket work - rewriting one reply, summarizing one ticket, classifying sentiment. These connectors are the cross-client, cross-system layer: analytics across thousands of tickets, ad-hoc questions, multi-system queries that join your PSA with your backup tool. The two complement each other; nothing gets ripped out. More in why msp-skills →.
Can I run this on Windows?
Yes. The PowerShell installer in “Install in 60 seconds” above is the same install path as macOS / Linux, just packaged for PowerShell. The CLI and MCP binaries are native Windows builds.
What does it cost?
The Skills, the CLI binaries, and the MCP servers are free. Apache-2.0 licensed - free to use commercially, free to fork. Servosity does not charge for API access required to run the Servosity connector. Other PSA, RMM, and backup vendors set their own API-access terms. You pay only for whichever AI agent you use (Claude, ChatGPT, Codex, etc.), and those are billed by your AI provider, not by us.
Can I try it on one client first?
Yes. Every command takes a client / company filter (e.g. halopsa-cli client card "Acme Corp", servosity-cli company show 4421). Read-tier commands change nothing; you can try them against your live system safely. For writes, each skill’s governance.md tags every command read / write / destructive with a recommended agent policy - the safe agent rule is to preview with --dry-run and require your approval before any mutation. The safe starting move is a read-only triage against one client, see the output, then widen scope.
What if my AI doesn’t speak MCP?
If your AI doesn’t yet support MCP and isn’t Claude Code / Codex CLI (which read SKILL.md directly), these connectors won’t fit yet. Every connector’s CLI binary still works standalone in a shell - you can pipe its output into anything. As more AI tools add MCP (the protocol is moving fast in 2026), they’ll work automatically without anything changing on our side.
Built and verified in Build Sessions
Every connector passes mechanical verification before it ships, and we build the next ones live with real MSPs running them against their own production systems in our free weekly Build Sessions. RSVP for the next one →.
Don’t see the system you need?
Tell us what to build next. Open an issue → and name the PSA, RMM, backup, security, or M365 tool. First time filing a GitHub issue? See the 90-second walkthrough →.
Standards. Conforms to the open Agent Skills spec (Anthropic, Dec 2025; 40+ agents). MCP-compatible - works with any MCP-capable agent including Hermes. OpenClaw-ready (frontmatter pre-wired, awaiting OpenClaw launch).
Built by Servosity. Maintained by Damien Stevens. Apache-2.0 licensed. Methodology: Compounding Teams.