// MCP TOKEN COST

The QuintaDB MCP server costs 24,830 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 QuintaDB MCP server's 116 tool definitions consume 24,830 tokens — 12% of a 200k context window, and 12× the median MCP server (2,145 tokens). A scoped grant exposing only the tools you use cuts that roughly in proportion.

MEASURED FROM SCHEMAS tiktoken o200k_base · rank #65 of 5,253 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 2.5%

Corpus context: QuintaDB ranks #65 of 5,253 measured MCP servers by definition cost. The median is 2,145 tokens, p90 is 11,440, and the heaviest (UnClick) is 147,411 — 74% 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 24,830 tokens go.

Each row is one tool definition as a tools/list entry — name, description and input schema — counted with o200k_base. Average: 214 tokens per tool.

Tool	Category	Tokens	% of server
create_field	Write	3,220	13.0%
update_field	Write	1,290	5.2%
add_dashboard_item	Write	867	3.5%
create_report	Write	774	3.1%
create_action_rule	Write	527	2.1%
create_formatting	Write	451	1.8%
create_team_group	Write	411	1.7%
bulk_update_records	Write	400	1.6%
add_portal_menu_item	Write	381	1.5%
create_map	Write	373	1.5%
update_portal_page	Write	365	1.5%
get_metric	Read	340	1.4%
create_record	Write	336	1.4%
update_portal_settings	Write	328	1.3%
create_calendar	Write	325	1.3%
create_chart	Write	325	1.3%
update_record	Write	322	1.3%
create_portal_group	Write	313	1.3%
place_media_on_site	Write	308	1.2%
create_portal	Write	297	1.2%
update_form_settings	Write	292	1.2%
create_email_template	Write	291	1.2%
create_saved_view	Write	256	1.0%
set_table_permission	Write	252	1.0%
import_records	Write	250	1.0%
set_column_size	Write	243	1.0%
update_portal_menu_item	Write	243	1.0%
upload_file_to_record	Write	241	1.0%
create_whatsapp_template	Write	237	1.0%
create_webhook	Write	235	0.9%
seed_records	Write	231	0.9%
delete_records_by_filter	Destructive	223	0.9%
add_user_to_team_group	Write	219	0.9%
configure_chat	Write	219	0.9%
create_schedule	Write	216	0.9%
create_reminder	Write	213	0.9%
add_user_to_portal_group	Write	211	0.8%
update_record_template	Write	210	0.8%
generate_ai_site	Write	200	0.8%
create_record_template	Write	199	0.8%
search_records	Read	197	0.8%
update_cell	Write	188	0.8%
set_module_permission	Write	185	0.7%
web_search	Read	181	0.7%
search_image	Read	180	0.7%
update_action_rule	Write	176	0.7%
get_records	Read	168	0.7%
attach_media_to_record	Write	164	0.7%
create_sms_template	Write	159	0.6%
create_telegram_template	Write	159	0.6%
get_portal_page	Read	156	0.6%
export_records	Read	156	0.6%
set_user_role	Write	156	0.6%
fetch_url	Read	150	0.6%
invite_portal_user	Write	150	0.6%
remove_user_from_portal_group	Destructive	147	0.6%
create_project	Write	136	0.5%
list_portal_users	Read	134	0.5%
reorder_fields	Write	134	0.5%
update_form	Write	130	0.5%
get_column_total	Read	129	0.5%
remove_user_from_team_group	Destructive	124	0.5%
delete_field	Destructive	123	0.5%
get_rel_id	Read	120	0.5%
delete_portal_group	Destructive	117	0.5%
create_folder	Write	117	0.5%
duplicate_record	Write	117	0.5%
delete_few_records	Destructive	111	0.4%
delete_portal_menu_item	Destructive	111	0.4%
get_form_fields	Read	110	0.4%
list_portal_menu_items	Read	109	0.4%
run_action	Execute	108	0.4%
list_templates	Read	108	0.4%
check_ai_site_status	Read	106	0.4%
generate_ai_project	Write	104	0.4%
delete_portal_user	Destructive	102	0.4%
delete_template	Destructive	101	0.4%
get_form_settings	Read	100	0.4%
list_portal_pages	Read	100	0.4%
list_saved_views	Read	99	0.4%
get_report_data	Read	98	0.4%
get_record	Read	97	0.4%
duplicate_form	Write	97	0.4%
get_field_types	Read	96	0.4%
delete_record	Destructive	95	0.4%
describe_project	Read	93	0.4%
get_reports	Read	93	0.4%
delete_form	Destructive	92	0.4%
delete_team_group	Destructive	92	0.4%
add_record_comment	Write	92	0.4%
get_quickstart_guide	Read	87	0.4%
list_portal_groups	Read	87	0.4%
create_form	Write	85	0.3%
list_folders	Read	82	0.3%
list_schedules	Read	81	0.3%
delete_saved_view	Destructive	79	0.3%
list_action_rules	Read	78	0.3%
list_webhooks	Read	77	0.3%
get_usage_stats	Read	76	0.3%
delete_reminder	Destructive	74	0.3%
delete_webhook	Destructive	74	0.3%
list_reminders	Read	74	0.3%
list_record_comments	Read	72	0.3%
update_project	Write	72	0.3%
delete_schedule	Destructive	70	0.3%
list_forms	Read	68	0.3%
list_user_roles	Read	67	0.3%
delete_project	Destructive	64	0.3%
list_projects	Read	64	0.3%
list_team_groups	Read	64	0.3%
list_portals	Read	59	0.2%
get_project	Read	58	0.2%
list_charts	Read	58	0.2%
list_maps	Read	54	0.2%
get_user_info	Read	53	0.2%
list_calendars	Read	52	0.2%

// CUT THE COST

Your agent uses a handful of these tools. It pays for all 116.

You don't need all 116 of those definitions in the window. PolicyLayer is an MCP gateway that sits in front of QuintaDB: 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 116 tools (no gateway)	24,830 tokens	—
3 granted tools	~642 tokens	−97%
5 granted tools	~1,070 tokens	−96%
10 granted tools	~2,141 tokens	−91%

The risk dividend: 17 of these 116 tools are critical-risk (destructive or financial) and cost 1,799 tokens (7% of the definition load). Block them — the recommended starter policy — and you reclaim that context before tuning anything else.

Create a free account and register QuintaDB — 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 QUINTADB TOKEN COST →

Instant setup, no code required.

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

//FAQ

QuintaDB token-cost questions.

How many tokens does the QuintaDB MCP server use?+

Its 116 tool definitions total 24,830 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 QuintaDB 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 QuintaDB's token usage?+

Expose fewer tools. A PolicyLayer grant scopes QuintaDB 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 642 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.

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 15-07-2026 from the PolicyLayer scan database over all 116 catalogued QuintaDB tools. Counts refresh with every site build.