// MCP TOKEN COST

The Yunxiao MCP server costs 24,443 tokens before the first call.

Every request your agent makes carries every tool definition this server exposes — context your code, documents and conversation can't use, mostly for tools the agent never calls. You don't need them all in the window, and you don't have to pay for them.

QUICK ANSWER The Yunxiao MCP server's 128 tool definitions consume 24,443 tokens — 12% of a 200k context window, and 12× the median MCP server (2,064 tokens). A scoped grant exposing only the tools you use cuts that roughly in proportion.

MEASURED FROM SCHEMAS tiktoken o200k_base · rank #50 of 3,875 measured servers · refreshed every build Method →

// CONTEXT WINDOW SHARE

What that costs 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 12%

1M WINDOW 2.4%

Corpus context: Yunxiao ranks #50 of 3,875 measured MCP servers by definition cost. The median is 2,064 tokens, p90 is 11,333, and the heaviest (SmartBear MCP) is 137,725 — 69% of a 200k window on its own. New to this? See MCP token cost and context window in the glossary.

Token cost is one axis. See the risk picture across your whole stack →

// PER-TOOL BREAKDOWN

Where the 24,443 tokens go.

Each row is one tool definition as a tools/list entry — name, description and input schema — counted with o200k_base. Average: 191 tokens per tool.

Tool	Category	Tokens	% of server
search_workitems	Read	805	3.3%
get_project_member_task_status	Read	415	1.7%
get_project_risk_dashboard	Read	386	1.6%
list_merge_requests	Read	326	1.3%
get_project_workitem_summary	Read	322	1.3%
get_project_overview	Read	317	1.3%
get_project_workitem_context	Read	317	1.3%
get_developer_usage	Read	309	1.3%
list_change_requests	Read	306	1.3%
get_repository_overview	Read	302	1.2%
list_pipelines	Read	292	1.2%
list_release_executions	Read	291	1.2%
search_organization_members	Read	288	1.2%
list_app_release_stage_runs	Read	285	1.2%
search_projects	Read	285	1.2%
list_change_request_comments	Read	282	1.2%
get_app_release_stage_overview	Read	279	1.1%
list_appstack_change_request_executions	Read	273	1.1%
get_project_workitem_detail	Read	270	1.1%
get_sprint_overview	Read	263	1.1%
get_team_workload_breakdown	Read	263	1.1%
list_commits	Read	259	1.1%
search_releases	Read	259	1.1%
get_project_workitem_board	Read	256	1.0%
get_branch_overview	Read	251	1.0%
list_pipeline_runs	Read	251	1.0%
get_organization_overview	Read	250	1.0%
get_department_usage	Read	249	1.0%
search_testcases	Read	242	1.0%
get_release_overview	Read	235	1.0%
get_change_request_overview	Read	234	1.0%
get_deployment_revision_info	Read	234	1.0%
get_my_project_workitems	Read	227	0.9%
list_artifacts	Read	227	0.9%
get_commit_overview	Read	224	0.9%
get_app_release_stage_pipeline_job_log	Read	223	0.9%
get_kubernetes_object_info	Read	222	0.9%
list_sprints	Read	222	0.9%
list_resource_instances	Read	221	0.9%
list_environments	Read	218	0.9%
list_repositories	Read	215	0.9%
search_app_templates	Read	211	0.9%
get_member_workload_trend	Read	209	0.9%
get_application_overview	Read	207	0.8%
get_pod_container_log	Read	206	0.8%
find_task_operation_log	Read	205	0.8%
list_knowledge_bases	Read	204	0.8%
list_change_order_job_logs	Read	198	0.8%
get_environment_overview	Read	197	0.8%
list_branches	Read	196	0.8%
get_work_item_type_overview	Read	195	0.8%
list_applications	Read	195	0.8%
list_kb_members	Read	195	0.8%
compare	Read	194	0.8%
get_app_release_stage_pipeline_run	Read	194	0.8%
list_kb_files	Read	193	0.8%
list_change_orders_by_origin	Read	192	0.8%
list_app_release_stage_exec_metadata	Read	191	0.8%
get_sprint_velocity	Read	190	0.8%
get_system_overview	Read	190	0.8%
search_app_tags	Read	189	0.8%
call_yunxiao_api	Execute	188	0.8%
get_pipeline_run_overview	Read	188	0.8%
list_change_order_versions	Read	184	0.8%
list_namespaces	Read	184	0.8%
list_package_repositories	Read	180	0.7%
get_pod_info	Read	175	0.7%
list_developer_members	Read	174	0.7%
get_app_orchestration	Read	173	0.7%
get_app_release_workflow_overview	Read	169	0.7%
get_pipeline_job_run_log	Read	167	0.7%
get_blocker_analysis	Read	166	0.7%
list_work_item_relation_work_item_types	Read	165	0.7%
get_change_order_overview	Read	164	0.7%
get_app_release_workflow_stage	Read	163	0.7%
get_artifact	Read	163	0.7%
list_users	Read	161	0.7%
get_file_blobs	Read	159	0.7%
list_organization_departments	Read	158	0.6%
list_files	Read	156	0.6%
list_work_item_types	Read	156	0.6%
get_appstack_change_request_audit_items	Read	155	0.6%
get_change_request_comment	Read	155	0.6%
get_organization_group_overview	Read	152	0.6%
get_pipeline_overview	Read	145	0.6%
get_workitem_status_timeline	Read	145	0.6%
list_project_members	Read	145	0.6%
list_change_request_patch_sets	Read	144	0.6%
get_merge_request	Read	139	0.6%
get_testcase	Read	139	0.6%
list_global_vars	Read	137	0.6%
list_resource_members	Read	136	0.6%
list_app_release_stage_briefs	Read	132	0.5%
list_organization_groups	Read	132	0.5%
list_pipeline_jobs_by_category	Read	132	0.5%
list_repository_members	Read	132	0.5%
get_organization_department_overview	Read	131	0.5%
get_machine_deploy_log	Read	130	0.5%
get_latest_orchestration	Read	128	0.5%
list_appstack_change_request_work_items	Read	126	0.5%
list_pipeline_relations	Read	126	0.5%
get_resource_instance	Read	125	0.5%
get_variable_group	Read	125	0.5%
list_systems	Read	124	0.5%
get_deploy_group	Read	123	0.5%
list_testcase_repositories	Read	123	0.5%
get_env_variable_groups	Read	121	0.5%
list_organization_members	Read	120	0.5%
list_all_work_item_types	Read	118	0.5%
describe_toolset	Read	116	0.5%
get_global_var	Read	116	0.5%
get_testcase_field_config	Read	114	0.5%
get_user_organizations	Read	105	0.4%
list_organizations	Read	104	0.4%
list_system_release_workflows	Read	103	0.4%
get_latest_pipeline_run	Read	102	0.4%
get_org_namespace	Read	102	0.4%
get_app_variable_groups	Read	98	0.4%
list_app_orchestration	Read	97	0.4%
list_directories	Read	96	0.4%
list_app_release_workflows	Read	95	0.4%
get_member_https_clone_username	Read	91	0.4%
list_app_extension_features	Read	88	0.4%
list_project_roles	Read	85	0.3%
list_organization_roles	Read	74	0.3%
get_user	Read	64	0.3%
get_current_organization_info	Read	58	0.2%
get_current_user	Read	56	0.2%

// CUT THE COST

Your agent uses a handful of these tools. It pays for all 128.

You don't need all 128 of those definitions in the window. PolicyLayer is an MCP gateway that sits in front of Yunxiao: only the tools you grant are exposed to the agent, the rest never load. A smaller window means a sharper agent — less noise when it picks a tool — and every request costs less:

Grant scope	Definition cost	Reduction
All 128 tools (no gateway)	24,443 tokens	—
3 granted tools	~573 tokens	−98%
5 granted tools	~955 tokens	−96%
10 granted tools	~1,910 tokens	−92%

Create a free account and register Yunxiao — nothing to install.
Grant only the tools you use — ungranted definitions never enter the context window.
Point your MCP client (Claude, Cursor, anything) at your gateway URL.

CUT YUNXIAO TOKEN COST →

Instant setup, no code required.

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

//FAQ

Yunxiao token-cost questions.

How many tokens does the Yunxiao MCP server use?+

Its 128 tool definitions total 24,443 tokens — 12% 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 Yunxiao 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 Yunxiao's token usage?+

Expose fewer tools. A PolicyLayer grant scopes Yunxiao 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 573 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 07-07-2026 from the PolicyLayer scan database over all 128 catalogued Yunxiao tools. Counts refresh with every site build.