# Running GroupDocs MCP servers on-premise: architecture and security model Run document signing for AI agents fully on-premise: local stdio transport, no external endpoints, no telemetry, and signing certificates that never leave your machine. Run document signing for AI agents **fully on-premise**: the GroupDocs.Signature MCP server uses local stdio transport with **no external endpoints, no inbound ports, and no telemetry** — suitable for regulated environments where neither documents nor signing certificates may leave the network. This page is the one to send your security reviewer. {{< alert style="info" >}} The commands and config snippets on this page are for the **.NET** build of the server — the only platform available today. Installation and client setup: [MCP server for .NET]({{< ref "signature/net/mcp/_index.md" >}}). Other platforms will expose the same tools with their own launch command; everything else on this page applies unchanged. {{< /alert >}} ## The architecture in one picture ```text +--------------+ +--------------------+ +------------------+ | AI client | stdio | MCP server process | reads / | local filesystem | | (Claude, VS | <-----> | (GroupDocs engine) | <-----> | storage / output | | Code, agent) | JSON-RPC | child process | writes | folders | +--------------+ +--------------------+ +------------------+ ``` * **Transport:** the AI client *starts the server as a child process* and communicates over standard input/output. The server never listens on a network socket — there is nothing to firewall, nothing to expose. * **Data path:** agent → local server → local filesystem. Documents and signed copies are read and written in the folders you configure; no document content is transmitted anywhere. * **Network use:** only at install time (pulling the package from nuget.org or the image from ghcr.io/docker.io). At runtime the server makes no outbound calls. In an air-gapped segment, pre-pull the image or pre-cache the package and pin the version. * **Telemetry:** none. The server does not phone home, and the engine processes documents in-process. ## The certificate question For a signing server this is the question a reviewer will ask first: **where does the private key go?** It stays on your machine. A `digital` signing call takes a certificate file path and a password; both are resolved locally by the local server process, the signing happens in-process, and neither the key nor the password is transmitted anywhere. There is no signing service in this architecture — no HSM proxy, no cloud key vault, no round trip. Two practical consequences: * **Protect the password like any other secret.** Prefer an environment variable over a literal in a committed client config; the same guidance as for [metered keys]({{< ref "signature/mcp/getting-started/licensing.md" >}}#keeping-the-private-key-out-of-committed-files) applies. * **Mount deliberately in Docker.** Give the container the certificate folder read-only, and only the document folders the agent should reach. What *does* travel to your model provider is the conversation: file names, decoded QR values the agent reports back, and any signature details it quotes. With a cloud-hosted model, assume anything the agent says out loud has left the building. Pair with a locally-hosted model and the whole loop stays inside the perimeter. ## Docker deployment inside the perimeter ```bash docker run --rm -i --user "$(id -u):$(id -g)" \ -v /srv/contracts:/data \ -v /srv/certs:/certs:ro \ -v /srv/licenses:/license:ro \ -e GROUPDOCS_MCP_STORAGE_PATH=/data \ -e GROUPDOCS_LICENSE_PATH=/license/GroupDocs.Signature.lic \ ghcr.io/groupdocs-signature/signature-net-mcp:26.9.0 ``` * Pin the tag (`:26.9.0`, not `:latest`) so a rebuild cannot change behaviour underneath you. * `--user "$(id -u):$(id -g)"` (Linux) lets the container, which runs as a non-root user, write into a folder you own. A JSON client config runs no shell to expand it — write the numbers: `"--user", "1000:1000"`. * Certificates and licence read-only; documents read-write. * Images are multi-arch (linux/amd64 + linux/arm64) and carry every native dependency. ## License management * **License file** — read from local disk by the local process. Fully offline; the right answer for air-gapped deployments. * **Metered (pay-per-use)** — reports *usage* to GroupDocs servers, so it needs outbound egress. Document content and certificates are never part of that report, but the connection must be allowed. Both are covered in [Licensing]({{< ref "signature/mcp/getting-started/licensing.md" >}}). ## What this fits — honestly **A good fit:** internal signing and verification workflows, compliance checks over document archives, regulated industries, air-gapped networks, and any team whose policy forbids sending documents or certificates to a third-party service. **Not what this is:** a signing service or a qualified trust service provider. It applies signatures with the certificate you give it, on one machine, for one client process. Whether a resulting signature satisfies a particular legal regime (eIDAS qualified, for example) depends on the certificate and the process around it, not on this server. ## FAQ **Does my certificate or its password leave the machine?** No. Both are read locally and used in-process. **Does any document content leave the machine?** Not from the server. What the agent quotes back to you travels in the conversation to your model provider. **Does it need internet at runtime?** No — only to fetch the package or image at install time, and only when metered licensing is enabled. **What ports does it open?** None. stdio only. **How do I prove that?** Run it and watch: no listening sockets, no outbound connections while signing. The [verification script]({{< ref "signature/net/mcp/troubleshooting.md" >}}#verifying-an-installation-end-to-end) performs a real handshake and a real engine call so you can observe exactly what happens. ## Read more * [3 architectures for AI document processing, and the one that keeps files inside your network](https://blog.groupdocs.com/total/on-premise-ai-document-processing-mcp/) — the same scenario on the GroupDocs blog, with worked examples and the trade-offs.