Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
Python agentaianthropicclaude-codecompressioncontext-engineeringcontext-windowcursorfastapilangchainllmmcpopenaiprompt-engineeringproxypythonragtoken-optimizationtokenstypescript
fix(dashboard): align token savings headline denominator (#1653)
Chris Yau646e70551436
diff --git a/headroom/dashboard/templates/dashboard.html b/headroom/dashboard/templates/dashboard.htmlindex 1aa01f8cb..389bf03be 100644--- a/headroom/dashboard/templates/dashboard.html+++ b/headroom/dashboard/templates/dashboard.html@@ -2565,26 +2565,14 @@// --- Headline savings percent ---//- // Active ratio (saved / attempted) is the right metric when we have- // a working denominator. Some backend-routed paths don't yet- // populate `proxy_attempted_tokens` consistently — in that case- // the active number reads 0% even when compression is happening,- // so fall back to the whole-request ratio (`proxy_savings_percent`)- // instead of showing a misleading zero. Issue #455.get headlineSavingsPercent() {- const attempted = this.stats.tokens?.proxy_attempted_tokens || 0;- const active = this.stats.tokens?.active_savings_percent || 0;- if (attempted > 0 && active > 0) return active;- return this.stats.tokens?.proxy_savings_percent || 0;+ return this.stats.tokens?.savings_percent+ ?? this.stats.tokens?.proxy_savings_percent+ ?? 0;},get headlineSavingsTitle() {- const attempted = this.stats.tokens?.proxy_attempted_tokens || 0;- const active = this.stats.tokens?.active_savings_percent || 0;- if (attempted > 0 && active > 0) {- return 'Of compressible tokens attempted';- }- return 'Of whole pre-compression request (attempted-tokens metric unavailable)';+ return 'Of total wire input tokens';},// --- Expandable Rows ---diff --git a/tests/test_dashboard_token_savings.py b/tests/test_dashboard_token_savings.pynew file mode 100644index 000000000..6c8b02569--- /dev/null+++ b/tests/test_dashboard_token_savings.py@@ -0,0 +1,27 @@+"""Regression tests for dashboard token savings copy."""++from __future__ import annotations++import re++from headroom.dashboard import get_dashboard_html+++def _getter_body(html: str, name: str) -> str:+ match = re.search(rf"get {name}\(\) \{{(?P<body>.*?)\n\s*\}},", html, re.S)+ assert match is not None+ return match.group("body")+++def test_token_savings_headline_uses_total_wire_denominator() -> None:+ html = get_dashboard_html()++ headline_body = _getter_body(html, "headlineSavingsPercent")+ assert "stats.tokens?.savings_percent" in headline_body+ assert "stats.tokens?.proxy_savings_percent" in headline_body+ assert "stats.tokens?.active_savings_percent" not in headline_body+ assert "stats.tokens?.proxy_attempted_tokens" not in headline_body++ title_body = _getter_body(html, "headlineSavingsTitle")+ assert "Of total wire input tokens" in title_body+ assert "Of compressible tokens attempted" not in title_body