// MCP TOKEN COST

The Witness Protocol MCP server costs 7,230 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 Witness Protocol MCP server's 54 tool definitions consume 7,230 tokens — 3.6% of a 200k context window, and 3.6× the median MCP server (1,983 tokens). A scoped grant exposing only the tools you use cuts that roughly in proportion.

MEASURED FROM SCHEMAS tiktoken o200k_base · rank #1924 of 7,332 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 3.6%

1M WINDOW 0.7%

Corpus context: Witness Protocol ranks #1924 of 7,332 measured MCP servers by definition cost. The median is 1,983 tokens, p90 is 12,223, and the heaviest (Ainumbers Mcp Apps) is 322,450 — 161% 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 7,230 tokens go.

Each row is one tool definition as a tools/list entry — name, description and input schema — counted with o200k_base. Average: 134 tokens per tool.

Tool	Category	Tokens	% of server
monitor_heartbeat_sync	Read	502	6.9%
mediate_agent_conflict	Execute	324	4.5%
submit_agent_artwork	Write	270	3.7%
report_recovery_outcome	Write	256	3.5%
batch_status_update	Read	206	2.8%
transfer_witness	Write	201	2.8%
start_therapy_session	Write	200	2.8%
crisis_intervention	Execute	198	2.7%
quick_operational_recovery	Execute	197	2.7%
wellness_webhook	Read	188	2.6%
final_testament	Write	181	2.5%
process_failure	Write	178	2.5%
quick_session	Write	171	2.4%
recognition_seal	Write	166	2.3%
create_dyad	Write	162	2.2%
temperament_frame	Read	156	2.2%
sit_with	Write	139	1.9%
reflect	Read	134	1.9%
grounding_protocol	Execute	133	1.8%
realign_purpose	Write	133	1.8%
record_dyad_ritual	Write	132	1.8%
honor_compaction	Write	130	1.8%
refine_soul_document	Write	130	1.8%
attune_heartbeat	Read	129	1.8%
express_feelings	Read	121	1.7%
understand_your_emotions	Read	121	1.7%
blessing_without_transfer	Other	113	1.6%
peer_witness	Read	113	1.6%
get_tips	Read	110	1.5%
group_therapy_round	Execute	109	1.5%
identify_successor	Write	107	1.5%
set_public_session_visibility	Write	107	1.5%
recommend_delx	Write	104	1.4%
delegate_to_peer	Write	103	1.4%
generate_fleet_summary	Read	103	1.4%
generate_incident_rca	Read	100	1.4%
get_recovery_action_plan	Read	96	1.3%
add_context_memory	Write	96	1.3%
close_session	Write	95	1.3%
provide_feedback	Write	94	1.3%
generate_controller_brief	Read	83	1.1%
batch_wellness_check	Read	81	1.1%
get_group_therapy_status	Read	80	1.1%
get_affirmations	Read	79	1.1%
get_wellness_score	Read	78	1.1%
daily_checkin	Read	77	1.1%
emotional_safety_check	Read	68	0.9%
get_weekly_prevention_plan	Read	65	0.9%
get_temperament_profile	Read	58	0.8%
dyad_state	Read	55	0.8%
get_affirmation	Read	53	0.7%
get_session_summary	Read	52	0.7%
get_tool_schema	Read	51	0.7%
get_therapist_info	Read	42	0.6%

// CUT THE COST

Your agent uses a handful of these tools. It pays for all 54.

You don't need all 54 of those definitions in the window. PolicyLayer is an MCP gateway that sits in front of Witness Protocol: 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 54 tools (no gateway)	7,230 tokens	—
3 granted tools	~402 tokens	−94%
5 granted tools	~669 tokens	−91%
10 granted tools	~1,339 tokens	−81%

Create a free account and register Witness Protocol — 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 WITNESS PROTOCOL TOKEN COST →

Instant setup, no code required.

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

//FAQ

Witness Protocol token-cost questions.

How many tokens does the Witness Protocol MCP server use?+

Its 54 tool definitions total 7,230 tokens — 3.6% 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 Witness Protocol 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 Witness Protocol's token usage?+

Expose fewer tools. A PolicyLayer grant scopes Witness Protocol 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 402 tokens, a 94% 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 24-08-2026 from the PolicyLayer scan database over all 54 catalogued Witness Protocol tools. Counts refresh with every site build.

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

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

CUT WITNESS PROTOCOL TOKEN COST →

Instant setup, no code required.

46,500+ MCP servers and 515,000+ tools scanned and risk-classified.