docs(benchmarks): measured context-overhead comparison vs official server#32
Merged
Conversation
…rver Reproducible benchmark: drives each MCP server through the stdio handshake, captures the real tools/list payload, counts tokens (o200k_base). Result — static tool-schema footprint at connection: official 17,163 tokens (24 tools) vs 422 (2 tools) = 97.5% smaller. Realistic multi-op sessions stay 85-95% lighter via on-demand notion_describe. Updates the README's '~90%' claim to the measured figure with a link to benchmarks/ for reproduction. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Replaces the README's unsourced "~90% less context overhead" with a measured, reproducible number — and ships the benchmark that produces it.
Result
Static tool-schema footprint (what every MCP client loads into the model's context at connection, and keeps there all session):
@notionhq/notion-mcp-server97.5% smaller — 40.7× less. Realistic multi-operation sessions stay 85–95% lighter, since operation schemas load on demand via
notion_describe(seebenchmarks/README.mdfor the scenario table and the honest worst case).Method
Drive each server through the MCP stdio handshake, capture the real
tools/listpayload, serialize each tool to the{name, description, input_schema}shape a client forwards to the model, count tokens withtiktokeno200k_base. No Notion token needed (schema listing is unauthenticated). Fully reproducible — commands inbenchmarks/README.md.Changes
benchmarks/— harness (list-tools.mjs,describe-all.mjs), token counter (count.py), andREADME.mdwith results + reproduction steps*.jsongitignoredbenchmarks/🤖 Generated with Claude Code