// MCP TOKEN COST

The Tailscale MCP server costs 10,834 tokens before the first call.

Connect Tailscale and its 89 tool definitions are loaded into the model's context on every request — 5.4% of a 200k window spent before your agent does anything.

QUICK ANSWER The Tailscale MCP server's tool definitions consume 10,834 tokens — 5.7× the median MCP server (1,905 tokens). A scoped grant exposing only the tools you use cuts that roughly in proportion.

MEASURED FROM SCHEMAS 89 tools · 10,834 tokens · 5.4% of 200k · 1.1% of 1M Method →

// CONTEXT WINDOW SHARE

What that buys before your agent starts working.

Tool definitions are overhead: they occupy context on every request and compete with your code, documents and conversation history for the same window.

200K WINDOW 5.4%

1M WINDOW 1.1%

Corpus context: Tailscale ranks #158 of 3,213 measured MCP servers by definition cost. The median is 1,905 tokens, p90 is 7,952, and the heaviest (Fusionauth) is 183,337 — 92% of a 200k window on its own.

// PER-TOOL BREAKDOWN

Where the 10,834 tokens go.

Each row is one tool definition as a tools/list entry — name, description and input schema — counted with o200k_base. Average: 122 tokens per tool.

Tool	Category	Tokens	% of server
tailscale_set_log_stream_config	Write	632	5.8%
tailscale_create_key	Write	570	5.3%
tailscale_set_contacts	Write	379	3.5%
tailscale_update_key	Write	314	2.9%
tailscale_update_tailnet_settings	Write	307	2.8%
tailscale_batch_update_posture_attributes	Write	269	2.5%
tailscale_create_posture_integration	Write	257	2.4%
tailscale_list_devices	Read	238	2.2%
tailscale_create_device_invite	Write	218	2.0%
tailscale_update_webhook	Write	207	1.9%
tailscale_set_device_ip	Write	204	1.9%
tailscale_set_device_posture_attribute	Write	186	1.7%
tailscale_create_webhook	Write	182	1.7%
tailscale_update_service	Write	176	1.6%
tailscale_create_user_invite	Write	175	1.6%
tailscale_set_dns_configuration	Write	173	1.6%
tailscale_update_acl	Write	171	1.6%
tailscale_preview_acl	Read	158	1.5%
tailscale_update_posture_integration	Write	158	1.5%
tailscale_set_devices_authorized	Write	156	1.4%
tailscale_update_split_dns	Write	151	1.4%
tailscale_set_device_routes	Write	139	1.3%
tailscale_set_split_dns	Write	137	1.3%
tailscale_get_network_flow_logs	Read	130	1.2%
tailscale_list_users	Read	123	1.1%
tailscale_get_audit_log	Read	122	1.1%
tailscale_set_device_tags	Write	121	1.1%
tailscale_delete_key	Destructive	120	1.1%
tailscale_list_services	Read	118	1.1%
tailscale_validate_aws_trust_policy	Read	116	1.1%
tailscale_set_service_device_approval	Write	112	1.0%
tailscale_update_device_key	Write	104	1.0%
tailscale_set_nameservers	Write	102	0.9%
tailscale_update_user_role	Write	102	0.9%
tailscale_delete_device_posture_attribute	Destructive	99	0.9%
tailscale_get_log_stream_status	Read	98	0.9%
tailscale_rename_device	Write	95	0.9%
tailscale_list_keys	Read	94	0.9%
tailscale_set_search_paths	Write	94	0.9%
tailscale_list_log_stream_configs	Read	93	0.9%
tailscale_rotate_webhook_secret	Write	92	0.8%
tailscale_delete_log_stream_config	Destructive	90	0.8%
tailscale_get_log_stream_config	Read	89	0.8%
tailscale_get_service_device_approval	Read	89	0.8%
tailscale_get_acl	Read	87	0.8%
tailscale_get_key	Read	85	0.8%
tailscale_suspend_user	Write	85	0.8%
tailscale_deauthorize_device	Read	84	0.8%
tailscale_delete_device	Destructive	81	0.7%
tailscale_resend_contact_verification	Read	81	0.7%
tailscale_delete_service	Destructive	80	0.7%
tailscale_delete_user	Destructive	78	0.7%
tailscale_get_device	Read	77	0.7%
tailscale_test_webhook	Read	77	0.7%
tailscale_get_service	Read	76	0.7%
tailscale_validate_acl	Read	76	0.7%
tailscale_delete_device_invite	Destructive	75	0.7%
tailscale_delete_user_invite	Destructive	75	0.7%
tailscale_delete_webhook	Destructive	74	0.7%
tailscale_create_aws_external_id	Write	73	0.7%
tailscale_restore_user	Write	73	0.7%
tailscale_set_dns_preferences	Write	73	0.7%
tailscale_get_device_posture_attributes	Read	72	0.7%
tailscale_expire_device	Read	70	0.6%
tailscale_accept_device_invite	Write	70	0.6%
tailscale_approve_user	Write	70	0.6%
tailscale_delete_posture_integration	Destructive	69	0.6%
tailscale_list_device_invites	Read	69	0.6%
tailscale_get_device_routes	Read	68	0.6%
tailscale_status	Read	68	0.6%
tailscale_get_dns_configuration	Read	67	0.6%
tailscale_get_posture_integration	Read	67	0.6%
tailscale_resend_device_invite	Write	67	0.6%
tailscale_resend_user_invite	Write	67	0.6%
tailscale_authorize_device	Read	65	0.6%
tailscale_get_device_invite	Read	65	0.6%
tailscale_get_user_invite	Read	65	0.6%
tailscale_list_service_hosts	Read	65	0.6%
tailscale_get_webhook	Read	64	0.6%
tailscale_get_user	Read	61	0.6%
tailscale_get_tailnet_settings	Read	57	0.5%
tailscale_get_dns_preferences	Read	53	0.5%
tailscale_get_contacts	Read	52	0.5%
tailscale_list_posture_integrations	Read	51	0.5%
tailscale_get_nameservers	Read	49	0.5%
tailscale_get_search_paths	Read	49	0.5%
tailscale_get_split_dns	Read	48	0.4%
tailscale_list_user_invites	Read	48	0.4%
tailscale_list_webhooks	Read	48	0.4%

// SCOPED GRANT

Most agents use a handful of these tools. They pay for all 89.

A PolicyLayer grant exposes only the tools you allow — ungranted definitions are filtered out of the tool list, so they never enter the context window. Estimates below assume typical-weight tools (122 tokens each).

Grant scope	Definition cost	Reduction
All 89 tools (no gateway)	10,834 tokens	—
3 granted tools	~365 tokens	−97%
5 granted tools	~609 tokens	−94%
10 granted tools	~1,217 tokens	−89%

Model your own stack in the token-cost calculator, or see the Tailscale policy for what a sensible grant looks like.

//FAQ

Tailscale token-cost questions.

How many tokens does the Tailscale MCP server use?+

Its 89 tool definitions total 10,834 tokens — 5.4% of a 200k context window — measured with tiktoken o200k_base over the serialised tools/list payload. Exact counts vary slightly by client and model.

Why does Tailscale consume tokens before I send a message?+

MCP clients load every connected server's tool definitions — name, description, and input schema — into the model's context so it knows what it can call. That payload is charged against your context window on every request, whether or not a tool is used.

How do I reduce Tailscale's token usage?+

Expose fewer tools. A PolicyLayer grant scopes Tailscale to only the tools you allow — ungranted definitions are filtered out of the tool list, so they never enter the context window. A grant of 3 typical tools costs roughly 365 tokens, a 97% reduction.

Does deferred tool loading fix this?+

Partially, in some clients. Claude Code defers MCP tool schemas behind a tool-search step by default, and VS Code has experimental grouping — but you still pay tokens per search and reload, and Cursor, Windsurf and Gemini CLI load definitions upfront. Reducing the exposed tool set cuts the cost in every client.

// METHOD & REVIEW

How these numbers were measured.

Serialisation

Each tool is serialised as a tools/list entry — name, description, input schema — from the schemas in the PolicyLayer scan database. Clients differ slightly in framing, so treat counts as close estimates.

Tokeniser

tiktoken o200k_base (GPT-4o/o-series). Anthropic's current tokeniser isn't published, so Claude's exact counts will differ; for English text and JSON schemas the totals are close enough to treat these as estimates.

Deferred loading

Some clients now defer schema loading (Claude Code's tool search; VS Code experimental grouping). You still pay per search and reload — and Cursor, Windsurf and Gemini CLI load everything upfront.

Computed 07-06-2026 from the PolicyLayer scan database over all 89 catalogued Tailscale tools. Counts refresh with every site build.

Expose only the tools you use — the rest never enter your context.

A PolicyLayer grant scopes Tailscale to the tools you actually allow. Ungranted definitions never load, and every call that does run is checked against policy first.

SCOPE YOUR GRANT →

Free to start. No card required.

4,600+ MCP servers and 31,000+ tools scanned and risk-classified.