hetzner-dns-mcp

A Model Context Protocol (MCP) server for managing DNS zones and records via the Hetzner Cloud API.

Lets MCP clients like Claude Code, Claude Desktop or Codex manage your Hetzner DNS: list, create, update and delete zones and RRSets (record sets), import/export zone files, manage protection and TTLs, and track asynchronous zone actions.

Note: This server targets the current DNS API that is part of the Hetzner Cloud API (api.hetzner.cloud). The legacy DNS API (dns.hetzner.com) was shut down in May 2026 and is not supported.

📖 Full documentation: hetzner-dns-mcp.ni-c.de

Requirements

• Node.js ≥ 22
• A Hetzner Cloud API token for the project that holds your DNS zones — create one in the Hetzner Cloud Console under your project → Security → API tokens. Use a read & write token for full functionality (a read-only token limits you to the read tools).

Configuration

Configuration is provided via environment variables:

Without a token the server still starts and lists its tools (so registries and inspectors can introspect it), but every tool call fails with setup instructions instead of reaching the API.

All three variables are deleted from the process environment once they have been read, so a later crash report or diagnostic dump cannot expose the token.

Installation

Claude Code

claude mcp add hetzner-dns -s user \
  -e HETZNER_API_TOKEN=your-token \
  -- npx -y hetzner-dns-mcp

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "hetzner-dns": {
      "command": "npx",
      "args": ["-y", "hetzner-dns-mcp"],
      "env": {
        "HETZNER_API_TOKEN": "your-token"
      }
    }
  }
}

Codex

Add to your ~/.codex/config.toml:

[mcp_servers.hetzner-dns]
command = "npx"
args = ["-y", "hetzner-dns-mcp"]
env = { HETZNER_API_TOKEN = "your-token" }

From source

git clone https://github.com/ni-c/hetzner-dns-mcp.git
cd hetzner-dns-mcp
npm install
npm run build
# then use `node /path/to/hetzner-dns-mcp/dist/index.js` as the command

Docker

Multi-arch images (amd64/arm64) are published to GHCR with an SBOM and build provenance:

docker run -i --rm -e HETZNER_API_TOKEN=your-token ghcr.io/ni-c/hetzner-dns-mcp

The image talks MCP over stdio, so clients need docker run -i (no port is exposed):

{
  "mcpServers": {
    "hetzner-dns": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "HETZNER_API_TOKEN",
        "ghcr.io/ni-c/hetzner-dns-mcp"
      ],
      "env": {
        "HETZNER_API_TOKEN": "your-token"
      }
    }
  }
}

Tools

Zones

RRSets (record sets)

Actions

Safety

Confirmation tokens. Every irreversible tool — delete_zone, delete_rrset, import_zonefile, set_records, remove_records, change_primary_nameservers, and change_zone_protection/change_rrset_protection when they remove protection — refuses its first call and returns a random, single-use token that is valid for five minutes. The second call must repeat the identical arguments and pass that token:

1. set_records(zone: "example.com", name: "www", type: "A", records: [{value: "198.51.100.1"}])
   → error: Refusing to replace the records … confirmToken: "3f9c…"
2. set_records(…same arguments…, confirmToken: "3f9c…")
   → executed

This is deliberately not a boolean the model can set on its own. The token exists only in a previous tool result, so an instruction hidden in a TXT record or a zone-file comment cannot manufacture one. Tokens for set_records, remove_records and import_zonefile are bound to a hash of the exact payload: a confirmation for ["198.51.100.1"] will not write ["198.51.100.66"].

The token is a guard rail, not a security boundary — the boundary is the permission prompt of your MCP host. Do not auto-approve these tools.

Untrusted upstream data. Everything the API returns is wrapped in an <untrusted-data> envelope, because record values, comments, labels and zone files are written by whoever controls the zone. Confirmation messages never quote that content; they report counts and TTLs only. Keys that look like secrets (tsig_key, token, secret, password) are redacted from results, oversized values are truncated, and HTML error pages from an intermediate proxy are dropped rather than pasted into the model's context.

Least privilege. Set HETZNER_READ_ONLY=true to register only the seven read tools — the write tools then do not exist on the protocol at all, rather than failing at call time. Combine it with a read-only Hetzner token for a genuinely read-only setup.

Other guarantees.

• Tools carry MCP annotations (readOnlyHint, destructiveHint, idempotentHint) so hosts can apply appropriate permission policies.
• HETZNER_API_BASE_URL must be an https URL (http is only accepted for localhost) and must not contain credentials; a warning is printed when a non-default host is configured, because the API token is sent there.
• Zone identifiers and RRSet names are validated against a strict character set and cannot escape the API path; requests never follow redirects, so the Bearer header cannot be replayed elsewhere.
• TSIG keys passed to create_zone/change_primary_nameservers become part of the conversation context and client transcripts — treat them as secrets and rotate them if in doubt.
• Hetzner-side resource protection is honored: protected zones/RRSets return an error with a hint to the corresponding change_*_protection tool.

Development

npm install
npm run build     # compile TypeScript to dist/
npm test          # run the vitest test suite
npm run lint      # eslint + prettier check

Releasing

1. Bump the version in package.json (npm version X.Y.Z --no-git-tag-version).
2. Rename the [Unreleased] section in CHANGELOG.md to the new version.
3. Commit, then tag and push: git tag -a vX.Y.Z -m "vX.Y.Z" && git push origin main vX.Y.Z.

The release workflow then runs the tests, publishes to npm via Trusted Publishing (with provenance), creates a GitHub release with the notes from the CHANGELOG and updates the entry in the official MCP Registry (io.github.ni-c/hetzner-dns-mcp, via GitHub OIDC).

License

MIT