Tailscale for AI agents
Tailscale runs a private mesh network connecting a team's devices and services, and network administrators use it to manage devices, routes, DNS, and access policy. Connect it once through Arc0, and your agent, or Claude, ChatGPT and Cursor, can use it through one MCP endpoint, limited to what each user approved.
What agents do in Tailscale.
List devices on the tailnet
Pull the current device roster and their routes as a read-only check before making any change.
Review configuration audit logs
Read recent audit log entries to see who changed access policy or device settings.
Confirm before changing device authorization
Set a device's authorization status only after review, since it changes what that device can reach.
10 Tailscale actions, graded by risk.
Every Tailscale action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.
read
7Look things up. Allowed by default.
- tailscale.get_deviceGet Device
- tailscale.list_usersList Users
- tailscale.list_device_routesList Device Routes
- tailscale.get_tailnet_settingsGet Tailnet Settings
- tailscale.list_tailnet_devicesList Tailnet Devices
- tailscale.get_dns_configurationGet DNS Configuration
- tailscale.list_configuration_audit_logsList Configuration Audit Logs
write
3Create and change things. Allow, or ask the user first.
- tailscale.set_device_tagsSet Device Tags
- tailscale.set_device_routesSet Device Routes
- tailscale.set_device_authorizationSet Device Authorization
destructive
0Delete, cancel or archive. Ask first, or deny outright.
- No destructive actions.
Tailscale in three steps.
- 01Your users connect TailscaleThey add their Tailscale api key on Arc0 Connect, under your brand. It goes straight into the vault.
- 02You set the rulesReads run, and writes like “set Device Tags” can wait for the user to approve.
- 03Any agent can actYour agent calls Tailscale through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.
await arc0.policies.set('tailscale', { read: 'allow', write: 'ask', // set_device_tags destructive: 'deny', }) # Claude Code: the same connection, one URL $ claude mcp add --transport http arc0 \ https://mcp.arc0.ai/u/u_8f2
How Tailscale connects.
Users add their Tailscale api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.
The same Tailscale connection serves your agent over MCP and your own backend over REST and the proxy, so a user connects once. How Arc0 handles credentials →
- AUTH
- API key
- CREDENTIALS
- Per-tenant encrypted vault
- MODEL SEES
- Results only, never credentials
- AUDIT LOG
- Every call, on every plan
Use Tailscale from any agent.
Tailscale and Arc0, answered.
Can I use Tailscale with Claude, ChatGPT or Cursor?
Yes. Connect Tailscale to Arc0 once, then add your Arc0 MCP URL to Claude, ChatGPT, Cursor, Claude Code or any other remote-MCP client. Each assistant only gets the Tailscale actions you allow.
How do users connect Tailscale?
Users add their Tailscale api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.
Which Tailscale actions can my agent take?
10 in total: 7 read, 3 write and 0 destructive, such as “set Device Tags”. Your policies decide which of them each agent may call.
Can I make my agent read-only in Tailscale?
Yes. Allow read actions and deny writes in the Tailscale policy. Your agent can still look things up, and any write it attempts is blocked and logged.
Can my own backend call Tailscale too?
Yes. The same Tailscale connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.
Plug Tailscale into your agent.
Your users connect Tailscale once, under your brand. Your agent gets 10 actions behind your policies, with every call on the record.
Free to build · MCP + REST · Audit log on every plan