// MCP TOKEN COST

The Dialogbrain MCP server costs 51,410 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 Dialogbrain MCP server's 220 tool definitions consume 51,410 tokens — 26% of a 200k context window, and 24× 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 #18 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 26%

1M WINDOW 5.1%

Corpus context: Dialogbrain ranks #18 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 51,410 tokens go.

Each row is one tool definition as a tools/list entry — name, description and input schema — counted with o200k_base. Average: 234 tokens per tool.

Tool	Category	Tokens	% of server
agents_update	Write	4,150	8.1%
documents_create	Write	2,543	4.9%
agents_trigger_create	Write	1,179	2.3%
videos_generate	Execute	1,033	2.0%
widgets_update	Write	1,001	1.9%
agents_simulate_inbound	Execute	863	1.7%
calls_send_to_meet	Execute	798	1.6%
calls_make	Execute	708	1.4%
calls_send_to_telegram_call	Execute	689	1.3%
agents_create	Write	541	1.1%
agents_trigger_update	Write	532	1.0%
messages_send	Write	517	1.0%
artifacts_create	Write	510	1.0%
integrations_add_endpoints	Write	500	1.0%
calls_list_history	Read	499	1.0%
agents_traces_list	Read	485	0.9%
reminder_set	Write	462	0.9%
knowledge_query	Read	448	0.9%
ai_filters_create	Write	422	0.8%
widgets_create	Write	419	0.8%
agent_handoff	Execute	417	0.8%
messages_read_history	Read	414	0.8%
notes_save	Write	408	0.8%
calendar_create_event	Write	406	0.8%
feedback_save	Write	392	0.8%
workbench_run_python	Execute	380	0.7%
contacts_update	Write	371	0.7%
group_list	Read	368	0.7%
linkedin_update_profile	Write	368	0.7%
ai_tags_create	Write	356	0.7%
kg_get_relationships	Read	355	0.7%
agents_ask	Execute	349	0.7%
integrations_set_auth	Write	349	0.7%
web_fetch	Read	343	0.7%
youtube_upload_video	Write	340	0.7%
agents_activity	Read	339	0.7%
contacts_find	Read	335	0.7%
search_threads	Read	329	0.6%
linkedin_search	Read	303	0.6%
background_run	Execute	290	0.6%
calls_agent_duel	Execute	289	0.6%
ai_filters_update	Write	289	0.6%
kg_find_entity	Read	286	0.6%
web_research	Read	282	0.5%
files_read	Read	274	0.5%
search_links	Read	273	0.5%
ai_tags_update	Write	270	0.5%
agents_trace_get	Read	269	0.5%
integrations_capture_session	Execute	268	0.5%
agents_set_integration	Write	260	0.5%
notes_recall	Read	255	0.5%
images_generate	Execute	251	0.5%
android_install_apk	Write	249	0.5%
search_files	Read	246	0.5%
images_search	Read	245	0.5%
notes_search	Read	243	0.5%
group_create	Write	242	0.5%
tasks_create	Write	238	0.5%
threads_delete	Destructive	234	0.5%
webhooks_configure_enrichment	Write	234	0.5%
integrations_remove_endpoints	Destructive	229	0.4%
group_promote_admin	Write	228	0.4%
contacts_capture_lead	Write	227	0.4%
calendar_check_availability	Read	226	0.4%
ai_filters_test	Read	220	0.4%
threads_update	Write	218	0.4%
messages_send_email	Write	216	0.4%
calls_set_translation_languages	Write	215	0.4%
messages_forward	Write	214	0.4%
group_search	Read	213	0.4%
calendar_update_event	Write	213	0.4%
calls_wait	Execute	211	0.4%
files_ingest	Write	206	0.4%
prompts_update	Write	203	0.4%
messages_edit	Write	202	0.4%
files_upload	Write	201	0.4%
web__local_search	Read	200	0.4%
ai_filters_list	Read	198	0.4%
calls_set_translation_language	Write	197	0.4%
search_messages	Read	195	0.4%
instagram_publish_media	Write	195	0.4%
contacts_add_channel	Write	192	0.4%
youtube_video_query	Read	191	0.4%
calls_meet_browser	Execute	189	0.4%
group_get_invite_link	Read	188	0.4%
tasks_list	Read	187	0.4%
linkedin_search_filters	Read	185	0.4%
youtube_update_video	Write	183	0.4%
browser_network_requests	Read	182	0.4%
calendar_list_events	Read	181	0.4%
agents_traces_stats	Read	180	0.4%
group_add	Write	178	0.3%
web_search	Read	175	0.3%
browser_type	Execute	174	0.3%
browser_file_upload	Write	173	0.3%
group_add_member	Write	173	0.3%
group_join	Write	170	0.3%
browser_take_screenshot	Execute	169	0.3%
tasks_update	Write	166	0.3%
workspace_switch	Execute	166	0.3%
agents_add_file	Write	164	0.3%
browser_console_messages	Read	162	0.3%
agents_list	Read	162	0.3%
agents_list_drafts	Read	160	0.3%
browser_select_option	Execute	159	0.3%
agents_update_from_template	Write	155	0.3%
agents_approve_draft	Write	154	0.3%
contacts_research	Read	154	0.3%
linkedin_raw_request	Execute	153	0.3%
ai_tags_add_to_thread	Write	152	0.3%
prompts_prompt_restore	Write	151	0.3%
agents_prompt_history	Read	150	0.3%
youtube_post_comment_reply	Write	150	0.3%
group_preview_messages	Read	149	0.3%
browser_tabs	Execute	149	0.3%
agent_silence	Other	148	0.3%
youtube_list_comments	Read	147	0.3%
android_swipe	Execute	147	0.3%
ai_tags_list	Read	144	0.3%
browser_evaluate	Execute	143	0.3%
files_get_base64	Read	143	0.3%
browser_wait_for	Execute	142	0.3%
prompts_prompt_history	Read	142	0.3%
artifacts_update	Write	141	0.3%
calls_get_transcript	Read	140	0.3%
agents_reject_draft	Write	140	0.3%
agents_prompt_restore	Write	139	0.3%
present_tab	Execute	137	0.3%
browser_handle_dialog	Execute	135	0.3%
channels_connect_telegram_bot	Write	135	0.3%
contacts_discover	Read	135	0.3%
artifacts_set_access	Write	134	0.3%
calls_hangup	Execute	132	0.3%
android_shell	Execute	131	0.3%
group_scan	Read	131	0.3%
browser_press_key	Execute	130	0.3%
artifacts_get	Read	128	0.2%
browser_snapshot	Read	127	0.2%
browser_attach_meet	Execute	126	0.2%
folders_create	Write	126	0.2%
calls_list_active	Read	124	0.2%
artifacts_export_pdf	Write	123	0.2%
browser_fill_form	Execute	120	0.2%
browser_drag	Execute	118	0.2%
browser_fill	Execute	118	0.2%
linkedin_invite	Write	118	0.2%
integrations_get_endpoints	Read	117	0.2%
ai_tags_remove_from_thread	Write	116	0.2%
agents_list_files	Read	116	0.2%
contacts_profile	Read	116	0.2%
system_sleep	Execute	114	0.2%
analytics_query	Read	114	0.2%
files_info	Read	113	0.2%
reminder_list	Read	112	0.2%
android_key	Execute	111	0.2%
agents_remove_file	Write	109	0.2%
vision_query	Read	109	0.2%
calendar_delete_event	Destructive	107	0.2%
collections_delete	Destructive	107	0.2%
agents_get	Read	107	0.2%
folders_delete	Destructive	105	0.2%
browser_open	Execute	104	0.2%
notes_delete	Destructive	103	0.2%
contacts_sync	Write	103	0.2%
youtube_moderate_comment	Write	103	0.2%
collections_add_file	Write	102	0.2%
youtube_list_videos	Read	101	0.2%
browser_resize	Execute	101	0.2%
widgets_get_embed_code	Read	100	0.2%
collections_list_files	Read	99	0.2%
agents_task_complete	Write	99	0.2%
collections_create	Write	98	0.2%
ai_filters_delete	Destructive	97	0.2%
browser_attach_identity	Write	94	0.2%
agents_list_integrations	Read	93	0.2%
collections_assign_agent	Write	92	0.2%
ai_tags_delete	Destructive	91	0.2%
instagram_list_media	Read	91	0.2%
job_update_context	Write	91	0.2%
collections_unassign_agent	Write	89	0.2%
messages_delete	Destructive	89	0.2%
browser_click	Execute	89	0.2%
instagram_update_media	Write	89	0.2%
job_escalate	Execute	85	0.2%
collections_list	Read	84	0.2%
linkedin_list_reactions	Read	84	0.2%
android_close_app	Execute	84	0.2%
job_complete	Write	84	0.2%
collections_remove_file	Write	83	0.2%
widgets_delete	Destructive	83	0.2%
linkedin_add_comment	Write	83	0.2%
youtube_delete_video	Destructive	82	0.2%
widgets_get	Read	82	0.2%
linkedin_get_profile	Read	81	0.2%
agents_trigger_delete	Destructive	80	0.2%
widgets_list	Read	78	0.2%
browser_hover	Execute	78	0.2%
reminder_cancel	Destructive	76	0.1%
tasks_delete	Destructive	76	0.1%
youtube_delete_comment	Destructive	76	0.1%
linkedin_list_invitations_sent	Read	75	0.1%
android_launch_app	Execute	74	0.1%
linkedin_list_connections	Read	74	0.1%
prompts_get	Read	74	0.1%
integrations_list	Read	72	0.1%
android_tap	Execute	71	0.1%
agents_delete	Destructive	67	0.1%
job_read_context	Read	66	0.1%
browser_navigate_back	Execute	65	0.1%
android_list_apps	Read	65	0.1%
linkedin_get_company	Read	65	0.1%
workspace_list	Read	64	0.1%
android_ui_dump	Read	59	0.1%
artifacts_list	Read	57	0.1%
prompts_list	Read	57	0.1%
workspace_current	Read	56	0.1%
browser_close	Execute	50	0.1%
android_type	Execute	47	0.1%
android_screenshot	Read	41	0.1%
android_current_app	Read	36	0.1%

// CUT THE COST

Your agent uses a handful of these tools. It pays for all 220.

You don't need all 220 of those definitions in the window. PolicyLayer is an MCP gateway that sits in front of Dialogbrain: 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 220 tools (no gateway)	51,410 tokens	—
3 granted tools	~701 tokens	−99%
5 granted tools	~1,168 tokens	−98%
10 granted tools	~2,337 tokens	−95%

The risk dividend: 16 of these 220 tools are critical-risk (destructive or financial) and cost 1,702 tokens (3% 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 Dialogbrain — 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 DIALOGBRAIN TOKEN COST →

Instant setup, no code required.

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

//FAQ

Dialogbrain token-cost questions.

How many tokens does the Dialogbrain MCP server use?+

Its 220 tool definitions total 51,410 tokens — 26% 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 Dialogbrain 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 Dialogbrain's token usage?+

Expose fewer tools. A PolicyLayer grant scopes Dialogbrain 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 701 tokens, a 99% 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 220 catalogued Dialogbrain tools. Counts refresh with every site build.