AnchorDesk is a modern, self-hosted service desk built for MSPs and IT teams. Two-way ticket sync with the PSAs you already use, device context from every RMM you run, email that becomes tickets, and customer self-service you control — while your data stays on your hardware.
Most helpdesks make you choose: their cloud, their workflow, their price per seat. AnchorDesk anchors everything to a database you own and treats every other system as a plug-in.
Self-hosted on your Docker or Kubernetes, with PostgreSQL as the single source of truth and a full audit trail on every change. MIT-licensed, no per-seat rent, no data hostage situations — export a printable ticket or the whole database whenever you like.
Tickets flow to and from ConnectWise and Jira with conflict handling a human approves. Devices from Tactical RMM, NinjaOne, and Datto merge into one record per machine, so every ticket knows the hardware behind it — and you can run scripts on it without leaving the ticket. Jira sync is beta, exercised against a live tenant; ConnectWise, NinjaOne, and Datto stay labelled alpha until a real tenant proves them — the integration list says exactly which is which.
Email becomes threaded tickets automatically. SLA clocks, manual deadlines, and automation rules keep promises visible and escalations hands-free. Live updates hit every screen instantly, it all works from a phone, and AI agents connect securely over MCP to work tickets like a technician.
Give customers one place to submit requests, follow the work, and tell you how support went — inside the same self-hosted AnchorDesk stack you own. You decide when the portal opens, who gets in, what they can see, and whether technicians who opt in can add a name and face to replies.
Customers can request access themselves, but registration never opens the door on its own. Administrators review each request and choose whether to approve or reject it; a matching email never grants access by itself, a company-domain match is only a review hint, and the portal is off by default.
Customers see their own tickets by default. Company-wide visibility is a deliberate admin choice and begins when their access takes effect; their own older requests stay visible. Replies remain “Support” by default. A name appears only when you enable named identities and the technician separately opts in; a photo appears only if they add one.
When customer feedback is enabled, customers can rate the support they received — positive, neutral, or negative, with an optional comment. Separately, you can let them mark a ticket they can access as solved. Their feedback cannot be edited or deleted. The Reports view tracks customer satisfaction company by company for the date range you choose.
Real screens from the current release — click any shot to zoom.
One Compose file brings up the database, backend, and web client. Connect your mail and integrations from the admin screen when you're ready.
Docker Compose or Kubernetes manifests, with prebuilt images on GHCR.
Point it at your mailboxes, PSA, and RMMs from Admin — credentials stay on your server.
Invite the team with SSO and MFA, set SLA policies, and let automation handle the busywork.
# The whole stack, one command git clone https://github.com/Spillers-Technology/AnchorDesk cd AnchorDesk docker compose up --build # Web on :5173 · API on :8060 · PostgreSQL on :5432
No seat counts, no license keys, no feature tiers. You pay only if you want the person who built it to help you run it.
Every feature, MIT-licensed, on your Docker or Kubernetes. Upgrades are documented per version, and backup and restore is written down and rehearsed, not assumed.
Setup, mailbox and RMM wiring, upgrades, and direct help, quoted per engagement. Tell us your team size and the tools you need connected at [email protected].
Run AnchorDesk in production and give regular, candid feedback, and deployment support is free. Especially valuable if you run ConnectWise Manage, NinjaOne, or Datto RMM — your tenant is what moves those adapters out of alpha.
Managed hosting is not offered yet. It waits on versioned schema migrations and a rehearsed restore path, because a host has to be able to promise both.
Free and open source under MIT. Run it yourself in minutes, or have Spillers Technology set it up and support it with you. Curious how it works? Read the documentation.