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
feat(dashboard): Cost Saved headline card; label per-row savings as message-only (#3353)
Tejas Chopra17b041235480
diff --git a/headroom/dashboard/templates/dashboard.html b/headroom/dashboard/templates/dashboard.htmlindex 81f8c8cb4..54c004919 100644--- a/headroom/dashboard/templates/dashboard.html+++ b/headroom/dashboard/templates/dashboard.html@@ -240,6 +240,26 @@</div></div>+ <!-- Cost Saved: single-pane dollars across every layer the+ session summary prices (compression, tool-schema+ deferral, prefix-cache net, extension-attributed+ dollars), split measured vs projected. -->+ <div class="bg-surface rounded-lg p-4 border border-border">+ <div class="text-xs text-gray-500 uppercase tracking-wide mb-1">Cost Saved</div>+ <div class="flex items-baseline gap-2">+ <span class="text-3xl font-light tabular-nums text-emerald-400"+ x-text="'$' + formatCurrency(stats.summary?.cost?.total_saved_usd || 0)"></span>+ <span class="text-sm text-emerald-400"+ x-text="(stats.summary?.cost?.savings_pct || 0).toFixed(1) + '%'"+ title="Total saved as a fraction of the without-Headroom baseline"></span>+ </div>+ <div class="mt-1 text-xs text-gray-500 leading-relaxed"+ x-text="'Measured $' + formatCurrency(stats.summary?.cost?.measured_saved_usd || 0) + ' · projected $' + formatCurrency(stats.summary?.cost?.projected_saved_usd || 0)"+ title="Measured: compression, tool-schema deferral, prefix-cache net. Projected: extension-attributed estimates such as model routing."></div>+ <div class="mt-1 text-xs text-gray-600 leading-relaxed"+ x-text="'$' + formatCurrency(stats.summary?.cost?.with_headroom_usd || 0) + ' spent vs $' + formatCurrency(stats.summary?.cost?.without_headroom_usd || 0) + ' baseline'"></div>+ </div>+<!-- Output Tokens Saved (counterfactual) --><div class="bg-surface rounded-lg p-4 border border-border"><div class="text-xs text-gray-500 uppercase tracking-wide mb-1">Output Tokens Saved</div>@@ -930,7 +950,8 @@<div class="px-4 py-3 font-medium">Model</div><div class="px-4 py-3 font-medium text-right">Input</div><div class="px-4 py-3 font-medium text-right">Output</div>- <div class="px-4 py-3 font-medium text-right">Saved</div>+ <div class="px-4 py-3 font-medium text-right"+ title="Message-compression share of this request only. Tool-schema deferral, cache discounts and routing savings are counted in the headline cards, not per row.">Msg Saved</div><div class="px-4 py-3 font-medium text-right">Latency</div></div><!-- Data rows -->