docs(benchmark): token-economy receipt + Phase 1 update#12
Merged
Conversation
Adds the canonical token-economy benchmark comparing three paths for
"open + read a GitHub README":
curl raw README 1 round-trip, 0 schema tax
browser-cli 1 compound, 0 schema tax
MCP chrome 3 round-trips, +schema tax (deferred mode)
v1 (2026-05-18) baseline:
curl 1,821 tokens · browser-cli 1,828 · MCP 2,560 (+41%)
Phase 1 update (2026-05-19) — same-day re-measurement after a
whitespace normalizer landed in the browser-cli sibling repo
(lib/commands/read.js):
curl 2,189 (README content grew between days)
browser-cli 1,896 (now wins curl by 13.4%, junk-line ratio
50% → 19%)
Framed as an architectural moat — "Why CLI Agents Have a 50-Year
Head Start" — not a token-shaving optimization. The moat is the
load-once command vocabulary of CLI tools vs MCP's per-tool schema
charge per turn.
Companion doc docs/contracts/browser-cli/USE_CASE_ROUTING.md
formalizes when each tool fits (content-addressable → curl;
auth/JS/provenance/policy → browser-cli) so readers don't
reduce the comparison to "just use curl."
Raw I/O captured byte-for-byte under raw/ for reproducibility.
Token proxy: chars / 3.8 (documented in count.py).
Co-Authored-By: Claude Opus 4.7 (1M context) <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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Adds the canonical token-economy benchmark comparing three paths for "open + read a GitHub README":
curl raw README 1 round-trip, 0 schema tax
browser-cli 1 compound, 0 schema tax
MCP chrome 3 round-trips, +schema tax (deferred mode)
v1 (2026-05-18) baseline:
curl 1,821 tokens · browser-cli 1,828 · MCP 2,560 (+41%)
Phase 1 update (2026-05-19) — same-day re-measurement after a whitespace normalizer landed in the browser-cli sibling repo (lib/commands/read.js):
curl 2,189 (README content grew between days)
browser-cli 1,896 (now wins curl by 13.4%, junk-line ratio
50% → 19%)
Framed as an architectural moat — "Why CLI Agents Have a 50-Year Head Start" — not a token-shaving optimization. The moat is the load-once command vocabulary of CLI tools vs MCP's per-tool schema charge per turn.
Companion doc docs/contracts/browser-cli/USE_CASE_ROUTING.md formalizes when each tool fits (content-addressable → curl; auth/JS/provenance/policy → browser-cli) so readers don't reduce the comparison to "just use curl."
Raw I/O captured byte-for-byte under raw/ for reproducibility. Token proxy: chars / 3.8 (documented in count.py).