// MCP TOKEN COST

The Countly MCP server costs 34,880 tokens before the first call.

Connect Countly and its 127 tool definitions are loaded into the model's context on every request — 17% of a 200k window spent before your agent does anything.

QUICK ANSWER The Countly MCP server's tool definitions consume 34,880 tokens — 32× the median MCP server (1,075 tokens). A scoped grant exposing only the tools you use cuts that roughly in proportion.

MEASURED FROM SCHEMAS 127 tools · 34,880 tokens · 17% of 200k · 3.5% 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 17%

1M WINDOW 3.5%

Corpus context: Countly ranks #12 of 1,659 measured MCP servers by definition cost. The median is 1,075 tokens, p90 is 6,119, and the heaviest (Fusionauth) is 183,337 — 92% of a 200k window on its own.

// PER-TOOL BREAKDOWN

Where the 34,880 tokens go.

Each row is one tool definition as a tools/list entry — name, description and input schema — counted with o200k_base. Average: 275 tokens per tool.

Tool	Category	Tokens	% of server
dashboards_widget_add	Write	3,345	9.6%
alerts_create	Write	1,339	3.8%
ab_experiments_create	Write	1,029	3.0%
formulas_run	Execute	990	2.8%
email_reports_core_create	Write	831	2.4%
email_reports_update	Write	739	2.1%
email_reports_dashboard_create	Write	607	1.7%
hooks_update	Write	582	1.7%
hooks_create	Write	573	1.6%
query_data	Read	545	1.6%
filtering_rules_create	Write	495	1.4%
funnels_create	Write	482	1.4%
app_users_update	Write	476	1.4%
cohorts_create	Write	449	1.3%
retention	Read	443	1.3%
formulas_save	Write	424	1.2%
databases_query	Read	403	1.2%
drill_bookmarks_create	Write	392	1.1%
dashboards_create	Write	389	1.1%
views_table	Read	378	1.1%
consents_history_search	Read	377	1.1%
dashboards_widget_update	Write	368	1.1%
filtering_rules_update	Write	364	1.0%
app_users_create	Write	360	1.0%
funnels_update	Write	347	1.0%
cohorts_update	Write	322	0.9%
dashboards_update	Write	319	0.9%
crash_groups_list	Read	312	0.9%
views_data	Read	308	0.9%
funnels_dropoff_users	Read	293	0.8%
session_frequency	Read	289	0.8%
notes_create	Write	287	0.8%
funnels_step_users	Read	273	0.8%
funnels_data	Read	271	0.8%
remote_config_parameters_update	Write	265	0.8%
consents_list	Read	264	0.8%
server_logs_contents	Read	264	0.8%
remote_config_parameters_add	Write	257	0.7%
times_of_day	Read	252	0.7%
remote_config_conditions_add	Write	248	0.7%
crashes_get	Read	245	0.7%
app_analytics_summary	Read	243	0.7%
cohorts_list	Read	242	0.7%
remote_config_conditions_update	Write	242	0.7%
session_durations	Read	239	0.7%
slipping_users	Read	227	0.7%
queriable_fields_list	Read	224	0.6%
collections_aggregate	Execute	223	0.6%
apps_update	Write	223	0.6%
events_create	Write	223	0.6%
drill_bookmarks_list	Read	222	0.6%
consents_stats	Read	221	0.6%
hooks_test	Read	221	0.6%
app_users_delete	Destructive	217	0.6%
cohorts_data	Read	215	0.6%
user_profiles_breakdown	Read	213	0.6%
dashboards_data	Read	209	0.6%
databases_document	Read	207	0.6%
user_loyalty	Read	207	0.6%
crashes_stats_get	Read	205	0.6%
user_profiles_query	Read	202	0.6%
notes_list	Read	201	0.6%
funnels_list	Read	197	0.6%
sdk_stats_get	Read	195	0.6%
crashes_comment_update	Write	195	0.6%
metadata_get	Read	193	0.6%
sdk_logs_list	Read	187	0.5%
datapoints_stats	Read	186	0.5%
email_reports_send	Write	176	0.5%
crashes_comment_delete	Destructive	174	0.5%
filtering_rules_toggle_status	Read	174	0.5%
email_reports_delete	Destructive	172	0.5%
email_reports_preview	Write	172	0.5%
hooks_delete	Destructive	171	0.5%
crashes_comment_add	Write	169	0.5%
ab_experiments_details	Read	165	0.5%
crashes_resolve	Write	164	0.5%
events_delete	Destructive	159	0.5%
filtering_rules_delete	Destructive	159	0.5%
user_profiles_get	Read	159	0.5%
crashes_unresolve	Destructive	155	0.4%
ab_experiments_start	Execute	155	0.4%
ab_experiments_list	Read	155	0.4%
server_logs_files_list	Read	155	0.4%
ab_experiments_stop	Execute	154	0.4%
hooks_list	Read	154	0.4%
ab_experiments_delete	Destructive	153	0.4%
remote_config_conditions_delete	Destructive	153	0.4%
remote_config_parameters_delete	Destructive	153	0.4%
apps_create	Write	153	0.4%
crashes_hide	Read	152	0.4%
email_reports_list	Read	150	0.4%
events_list	Read	150	0.4%
crashes_show	Read	148	0.4%
drill_bookmarks_delete	Destructive	143	0.4%
formulas_delete	Destructive	143	0.4%
alerts_delete	Destructive	141	0.4%
live_last_day	Read	140	0.4%
cohorts_delete	Destructive	139	0.4%
sdk_config_get	Read	138	0.4%
apps_reset	Destructive	137	0.4%
live_last_hour	Read	137	0.4%
funnels_delete	Destructive	135	0.4%
live_last_30_days	Read	133	0.4%
datapoints_punch_card	Read	131	0.4%
live_users	Read	130	0.4%
databases_stats	Read	127	0.4%
filtering_rules_list	Read	127	0.4%
remote_configs_list	Read	126	0.4%
apps_delete	Destructive	124	0.4%
datapoints_top_apps	Read	124	0.4%
live_metrics	Read	123	0.4%
formulas_list	Read	122	0.3%
live_overall	Read	121	0.3%
collections_indexes	Read	117	0.3%
dashboards_widget_remove	Destructive	116	0.3%
alerts_list	Read	116	0.3%
apps_get_by_name	Read	105	0.3%
dashboards_list	Read	105	0.3%
notes_delete	Destructive	88	0.3%
apps_list	Read	85	0.2%
dashboards_delete	Destructive	78	0.2%
databases_list	Read	70	0.2%
get_plugins	Read	67	0.2%
dashboard_users	Read	65	0.2%
ping	Read	56	0.2%
get_version	Read	42	0.1%

// SCOPED GRANT

Most agents use a handful of these tools. They pay for all 127.

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 (275 tokens each).

Grant scope	Definition cost	Reduction
All 127 tools (no gateway)	34,880 tokens	—
3 granted tools	~824 tokens	−98%
5 granted tools	~1,373 tokens	−96%
10 granted tools	~2,746 tokens	−92%

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

//FAQ

Countly token-cost questions.

How many tokens does the Countly MCP server use?+

Its 127 tool definitions total 34,880 tokens — 17% 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 Countly 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 Countly's token usage?+

Expose fewer tools. A PolicyLayer grant scopes Countly 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 824 tokens, a 98% 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.

01

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.

02

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.

03

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 05-06-2026 from the PolicyLayer scan database over all 127 catalogued Countly tools. Counts refresh with every site build.