fix(ui): use current context usage in Control UI

vincentkoc · web-flow · commit ebb08dc70e1f · 2026-04-25T00:54:33.000-07:00
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -14,6 +14,7 @@ Docs: https://docs.openclaw.ai
 ### Fixes
 
 - OpenAI/Codex image generation: canonicalize legacy `openai-codex.baseUrl` values such as `https://chatgpt.com/backend-api` to the Codex Responses backend before calling `gpt-image-2`, matching the chat transport. Fixes #71460.
+- Control UI: make `/usage` use the fresh context snapshot for context percentage, and include cache-write tokens in the Usage overview cache-hit denominator. Fixes #47885. Thanks @imwyvern and @Ante042.
 - Telegram/webhook: acknowledge validated webhook updates before running bot middleware, keeping slow agent turns from tripping Telegram delivery retries while preserving per-chat processing lanes. Fixes #71392.
 - MCP: retire one-shot embedded bundled MCP runtimes at run end, skip bundle-MCP startup when a runtime tool allowlist cannot reach bundle-MCP tools, and add `mcp.sessionIdleTtlMs` idle eviction for leaked session runtimes. Fixes #71106, #71110, #70389, and #70808.
 - MCP/config reload: hot-apply `mcp.*` changes by disposing cached session MCP runtimes, and dispose bundled MCP runtimes during gateway shutdown so removed `mcp.servers` entries reap child processes promptly. Fixes #60656.
diff --git a/ui/src/i18n/.i18n/de.meta.json b/ui/src/i18n/.i18n/de.meta.json
@@ -1,10 +1,10 @@
 {
   "fallbackKeys": [],
-  "generatedAt": "2026-04-24T20:26:59.540Z",
+  "generatedAt": "2026-04-25T07:32:16.675Z",
   "locale": "de",
-  "model": "gpt-5.5",
+  "model": "gpt-5.4",
   "provider": "openai",
-  "sourceHash": "33cba33627744c6bb03182f53b02a9a2640272ef0fa3d039ce52723a96f5099e",
+  "sourceHash": "2af900ae253948aab69216e38e0fce2dfde89801d178dee0ebb8dd28df2e11ef",
   "totalKeys": 734,
   "translatedKeys": 734,
   "workflow": 1
diff --git a/ui/src/i18n/.i18n/es.meta.json b/ui/src/i18n/.i18n/es.meta.json
@@ -1,10 +1,10 @@
 {
   "fallbackKeys": [],
-  "generatedAt": "2026-04-24T20:27:55.136Z",
+  "generatedAt": "2026-04-25T07:32:18.681Z",
   "locale": "es",
-  "model": "gpt-5.5",
+  "model": "gpt-5.4",
   "provider": "openai",
-  "sourceHash": "33cba33627744c6bb03182f53b02a9a2640272ef0fa3d039ce52723a96f5099e",
+  "sourceHash": "2af900ae253948aab69216e38e0fce2dfde89801d178dee0ebb8dd28df2e11ef",
   "totalKeys": 734,
   "translatedKeys": 734,
   "workflow": 1
diff --git a/ui/src/i18n/.i18n/fr.meta.json b/ui/src/i18n/.i18n/fr.meta.json
@@ -1,10 +1,10 @@
 {
   "fallbackKeys": [],
-  "generatedAt": "2026-04-24T20:29:33.727Z",
+  "generatedAt": "2026-04-25T07:32:27.034Z",
   "locale": "fr",
-  "model": "gpt-5.5",
+  "model": "gpt-5.4",
   "provider": "openai",
-  "sourceHash": "33cba33627744c6bb03182f53b02a9a2640272ef0fa3d039ce52723a96f5099e",
+  "sourceHash": "2af900ae253948aab69216e38e0fce2dfde89801d178dee0ebb8dd28df2e11ef",
   "totalKeys": 734,
   "translatedKeys": 734,
   "workflow": 1
diff --git a/ui/src/i18n/.i18n/id.meta.json b/ui/src/i18n/.i18n/id.meta.json
@@ -1,10 +1,10 @@
 {
   "fallbackKeys": [],
-  "generatedAt": "2026-04-24T20:30:08.920Z",
+  "generatedAt": "2026-04-25T07:32:34.367Z",
   "locale": "id",
-  "model": "gpt-5.5",
+  "model": "gpt-5.4",
   "provider": "openai",
-  "sourceHash": "33cba33627744c6bb03182f53b02a9a2640272ef0fa3d039ce52723a96f5099e",
+  "sourceHash": "2af900ae253948aab69216e38e0fce2dfde89801d178dee0ebb8dd28df2e11ef",
   "totalKeys": 734,
   "translatedKeys": 734,
   "workflow": 1
diff --git a/ui/src/i18n/.i18n/ja-JP.meta.json b/ui/src/i18n/.i18n/ja-JP.meta.json
@@ -1,10 +1,10 @@
 {
   "fallbackKeys": [],
-  "generatedAt": "2026-04-24T20:29:06.508Z",
+  "generatedAt": "2026-04-25T07:32:20.713Z",
   "locale": "ja-JP",
-  "model": "gpt-5.5",
+  "model": "gpt-5.4",
   "provider": "openai",
-  "sourceHash": "33cba33627744c6bb03182f53b02a9a2640272ef0fa3d039ce52723a96f5099e",
+  "sourceHash": "2af900ae253948aab69216e38e0fce2dfde89801d178dee0ebb8dd28df2e11ef",
   "totalKeys": 734,
   "translatedKeys": 734,
   "workflow": 1
diff --git a/ui/src/i18n/.i18n/ko.meta.json b/ui/src/i18n/.i18n/ko.meta.json
@@ -1,10 +1,10 @@
 {
   "fallbackKeys": [],
-  "generatedAt": "2026-04-24T20:29:26.571Z",
+  "generatedAt": "2026-04-25T07:32:24.032Z",
   "locale": "ko",
-  "model": "gpt-5.5",
+  "model": "gpt-5.4",
   "provider": "openai",
-  "sourceHash": "33cba33627744c6bb03182f53b02a9a2640272ef0fa3d039ce52723a96f5099e",
+  "sourceHash": "2af900ae253948aab69216e38e0fce2dfde89801d178dee0ebb8dd28df2e11ef",
   "totalKeys": 734,
   "translatedKeys": 734,
   "workflow": 1
diff --git a/ui/src/i18n/.i18n/pl.meta.json b/ui/src/i18n/.i18n/pl.meta.json
@@ -1,10 +1,10 @@
 {
   "fallbackKeys": [],
-  "generatedAt": "2026-04-24T20:30:13.404Z",
+  "generatedAt": "2026-04-25T07:32:38.432Z",
   "locale": "pl",
-  "model": "gpt-5.5",
+  "model": "gpt-5.4",
   "provider": "openai",
-  "sourceHash": "33cba33627744c6bb03182f53b02a9a2640272ef0fa3d039ce52723a96f5099e",
+  "sourceHash": "2af900ae253948aab69216e38e0fce2dfde89801d178dee0ebb8dd28df2e11ef",
   "totalKeys": 734,
   "translatedKeys": 734,
   "workflow": 1
diff --git a/ui/src/i18n/.i18n/pt-BR.meta.json b/ui/src/i18n/.i18n/pt-BR.meta.json
@@ -1,10 +1,10 @@
 {
   "fallbackKeys": [],
-  "generatedAt": "2026-04-24T20:27:01.151Z",
+  "generatedAt": "2026-04-25T07:32:14.907Z",
   "locale": "pt-BR",
-  "model": "gpt-5.5",
+  "model": "gpt-5.4",
   "provider": "openai",
-  "sourceHash": "33cba33627744c6bb03182f53b02a9a2640272ef0fa3d039ce52723a96f5099e",
+  "sourceHash": "2af900ae253948aab69216e38e0fce2dfde89801d178dee0ebb8dd28df2e11ef",
   "totalKeys": 734,
   "translatedKeys": 734,
   "workflow": 1
diff --git a/ui/src/i18n/.i18n/th.meta.json b/ui/src/i18n/.i18n/th.meta.json
@@ -1,10 +1,10 @@
 {
   "fallbackKeys": [],
-  "generatedAt": "2026-04-24T20:30:41.102Z",
+  "generatedAt": "2026-04-25T07:32:40.554Z",
   "locale": "th",
-  "model": "gpt-5.5",
+  "model": "gpt-5.4",
   "provider": "openai",
-  "sourceHash": "33cba33627744c6bb03182f53b02a9a2640272ef0fa3d039ce52723a96f5099e",
+  "sourceHash": "2af900ae253948aab69216e38e0fce2dfde89801d178dee0ebb8dd28df2e11ef",
   "totalKeys": 734,
   "translatedKeys": 734,
   "workflow": 1
diff --git a/ui/src/i18n/.i18n/tr.meta.json b/ui/src/i18n/.i18n/tr.meta.json
@@ -1,10 +1,10 @@
 {
   "fallbackKeys": [],
-  "generatedAt": "2026-04-24T20:29:55.552Z",
+  "generatedAt": "2026-04-25T07:32:30.128Z",
   "locale": "tr",
-  "model": "gpt-5.5",
+  "model": "gpt-5.4",
   "provider": "openai",
-  "sourceHash": "33cba33627744c6bb03182f53b02a9a2640272ef0fa3d039ce52723a96f5099e",
+  "sourceHash": "2af900ae253948aab69216e38e0fce2dfde89801d178dee0ebb8dd28df2e11ef",
   "totalKeys": 734,
   "translatedKeys": 734,
   "workflow": 1
diff --git a/ui/src/i18n/.i18n/uk.meta.json b/ui/src/i18n/.i18n/uk.meta.json
@@ -1,10 +1,10 @@
 {
   "fallbackKeys": [],
-  "generatedAt": "2026-04-24T20:30:10.003Z",
+  "generatedAt": "2026-04-25T07:32:32.725Z",
   "locale": "uk",
-  "model": "gpt-5.5",
+  "model": "gpt-5.4",
   "provider": "openai",
-  "sourceHash": "33cba33627744c6bb03182f53b02a9a2640272ef0fa3d039ce52723a96f5099e",
+  "sourceHash": "2af900ae253948aab69216e38e0fce2dfde89801d178dee0ebb8dd28df2e11ef",
   "totalKeys": 734,
   "translatedKeys": 734,
   "workflow": 1
diff --git a/ui/src/i18n/.i18n/zh-CN.meta.json b/ui/src/i18n/.i18n/zh-CN.meta.json
@@ -1,10 +1,10 @@
 {
   "fallbackKeys": [],
-  "generatedAt": "2026-04-24T20:26:33.969Z",
+  "generatedAt": "2026-04-25T07:32:11.950Z",
   "locale": "zh-CN",
-  "model": "gpt-5.5",
+  "model": "gpt-5.4",
   "provider": "openai",
-  "sourceHash": "33cba33627744c6bb03182f53b02a9a2640272ef0fa3d039ce52723a96f5099e",
+  "sourceHash": "2af900ae253948aab69216e38e0fce2dfde89801d178dee0ebb8dd28df2e11ef",
   "totalKeys": 734,
   "translatedKeys": 734,
   "workflow": 1
diff --git a/ui/src/i18n/.i18n/zh-TW.meta.json b/ui/src/i18n/.i18n/zh-TW.meta.json
@@ -1,10 +1,10 @@
 {
   "fallbackKeys": [],
-  "generatedAt": "2026-04-24T20:27:07.043Z",
+  "generatedAt": "2026-04-25T07:32:13.730Z",
   "locale": "zh-TW",
-  "model": "gpt-5.5",
+  "model": "gpt-5.4",
   "provider": "openai",
-  "sourceHash": "33cba33627744c6bb03182f53b02a9a2640272ef0fa3d039ce52723a96f5099e",
+  "sourceHash": "2af900ae253948aab69216e38e0fce2dfde89801d178dee0ebb8dd28df2e11ef",
   "totalKeys": 734,
   "translatedKeys": 734,
   "workflow": 1
diff --git a/ui/src/i18n/locales/en.ts b/ui/src/i18n/locales/en.ts
@@ -535,7 +535,8 @@ export const en: TranslationMap = {
       errorHint: "Error rate = errors / total messages. Lower is better.",
       avgSession: "avg session",
       cacheHitRate: "Cache Hit Rate",
-      cacheHint: "Cache hit rate = cache read / (input + cache read). Higher is better.",
+      cacheHint:
+        "Cache hit rate = cache read / (input + cache read + cache write). Higher is better.",
       cached: "cached",
       prompt: "prompt",
       calls: "calls",
diff --git a/ui/src/ui/chat/slash-command-executor.node.test.ts b/ui/src/ui/chat/slash-command-executor.node.test.ts
@@ -488,7 +488,70 @@ describe("executeSlashCommand directives", () => {
     );
 
     expect(result.content).toBe(
-      "**Session Usage**\nInput: **1.2k** tokens\nOutput: **300** tokens\nTotal: **1.5k** tokens\nContext: **30%** of 4k\nModel: `gpt-4.1-mini`",
+      "**Session Usage**\nInput: **1.2k** tokens\nOutput: **300** tokens\nTotal: **1.5k** tokens\nContext: **38%** of 4k\nModel: `gpt-4.1-mini`",
+    );
+    expect(request).toHaveBeenNthCalledWith(1, "sessions.list", {});
+  });
+
+  it("keeps /usage context hidden when the context snapshot is stale", async () => {
+    const request = vi.fn(async (method: string, _payload?: unknown) => {
+      if (method === "sessions.list") {
+        return {
+          sessions: [
+            row("agent:main:main", {
+              model: "gpt-4.1-mini",
+              inputTokens: 1200,
+              outputTokens: 300,
+              totalTokens: 1500,
+              totalTokensFresh: false,
+              contextTokens: 4000,
+            }),
+          ],
+        };
+      }
+      throw new Error(`unexpected method: ${method}`);
+    });
+
+    const result = await executeSlashCommand(
+      { request } as unknown as GatewayBrowserClient,
+      "agent:main:main",
+      "usage",
+      "",
+    );
+
+    expect(result.content).toBe(
+      "**Session Usage**\nInput: **1.2k** tokens\nOutput: **300** tokens\nTotal: **~1.5k** tokens\nModel: `gpt-4.1-mini`",
+    );
+    expect(request).toHaveBeenNthCalledWith(1, "sessions.list", {});
+  });
+
+  it("uses the context snapshot for /usage while preserving cumulative total display", async () => {
+    const request = vi.fn(async (method: string, _payload?: unknown) => {
+      if (method === "sessions.list") {
+        return {
+          sessions: [
+            row("agent:main:main", {
+              model: "gpt-4.1-mini",
+              inputTokens: 1200,
+              outputTokens: 300,
+              totalTokens: 1250,
+              contextTokens: 4000,
+            }),
+          ],
+        };
+      }
+      throw new Error(`unexpected method: ${method}`);
+    });
+
+    const result = await executeSlashCommand(
+      { request } as unknown as GatewayBrowserClient,
+      "agent:main:main",
+      "usage",
+      "",
+    );
+
+    expect(result.content).toBe(
+      "**Session Usage**\nInput: **1.2k** tokens\nOutput: **300** tokens\nTotal: **1.5k** tokens\nContext: **31%** of 4k\nModel: `gpt-4.1-mini`",
     );
     expect(request).toHaveBeenNthCalledWith(1, "sessions.list", {});
   });
diff --git a/ui/src/ui/chat/slash-command-executor.ts b/ui/src/ui/chat/slash-command-executor.ts
@@ -377,17 +377,30 @@ async function executeUsage(
     if (!session) {
       return { content: "No active session." };
     }
-    const input = session.inputTokens ?? 0;
-    const output = session.outputTokens ?? 0;
-    const total = session.totalTokens ?? input + output;
+    const hasInputTokens = Number.isFinite(session.inputTokens);
+    const hasOutputTokens = Number.isFinite(session.outputTokens);
+    const input = hasInputTokens ? (session.inputTokens ?? 0) : 0;
+    const output = hasOutputTokens ? (session.outputTokens ?? 0) : 0;
+    const cumulativeTotal = hasInputTokens || hasOutputTokens ? input + output : null;
+    const contextSnapshotTotal = Number.isFinite(session.totalTokens)
+      ? (session.totalTokens ?? null)
+      : cumulativeTotal;
+    const totalTokensFresh = session.totalTokensFresh !== false;
     const ctx = session.contextTokens ?? 0;
-    const pct = ctx > 0 ? Math.round((input / ctx) * 100) : null;
+    const pct =
+      contextSnapshotTotal !== null && totalTokensFresh && ctx > 0
+        ? Math.round((contextSnapshotTotal / ctx) * 100)
+        : null;
+    const totalDisplay =
+      cumulativeTotal === null
+        ? "n/a"
+        : `${totalTokensFresh ? "" : "~"}${fmtTokens(cumulativeTotal)}`;
 
     const lines = [
       "**Session Usage**",
       `Input: **${fmtTokens(input)}** tokens`,
       `Output: **${fmtTokens(output)}** tokens`,
-      `Total: **${fmtTokens(total)}** tokens`,
+      `Total: **${totalDisplay}** tokens`,
     ];
     if (pct !== null) {
       lines.push(`Context: **${pct}%** of ${fmtTokens(ctx)}`);
diff --git a/ui/src/ui/views/usage-render-overview.test.ts b/ui/src/ui/views/usage-render-overview.test.ts
@@ -0,0 +1,69 @@
+/* @vitest-environment jsdom */
+
+import { render } from "lit";
+import { describe, expect, it } from "vitest";
+import { renderUsageInsights } from "./usage-render-overview.ts";
+import type { UsageAggregates, UsageTotals } from "./usageTypes.ts";
+
+const totals: UsageTotals = {
+  input: 100,
+  output: 40,
+  cacheRead: 300,
+  cacheWrite: 600,
+  totalTokens: 1040,
+  totalCost: 0,
+  inputCost: 0,
+  outputCost: 0,
+  cacheReadCost: 0,
+  cacheWriteCost: 0,
+  missingCostEntries: 0,
+};
+
+const aggregates = {
+  messages: {
+    total: 4,
+    user: 2,
+    assistant: 2,
+    toolCalls: 0,
+    toolResults: 0,
+    errors: 0,
+  },
+  tools: {
+    totalCalls: 0,
+    uniqueTools: 0,
+    tools: [],
+  },
+  byModel: [],
+  byProvider: [],
+  byAgent: [],
+  byChannel: [],
+  daily: [],
+} as unknown as UsageAggregates;
+
+describe("renderUsageInsights", () => {
+  it("includes cache writes in cache-hit-rate denominator", () => {
+    const container = document.createElement("div");
+
+    render(
+      renderUsageInsights(
+        totals,
+        aggregates,
+        {
+          durationSumMs: 0,
+          durationCount: 0,
+          avgDurationMs: 0,
+          errorRate: 0,
+        },
+        false,
+        [],
+        1,
+        1,
+      ),
+      container,
+    );
+
+    expect(container.textContent).toContain("30.0%");
+    expect(container.textContent).toContain("300 cached");
+    expect(container.textContent).toContain("1.0K prompt");
+  });
+});
diff --git a/ui/src/ui/views/usage-render-overview.ts b/ui/src/ui/views/usage-render-overview.ts
@@ -497,7 +497,7 @@ function renderUsageInsights(
     ? Math.round(totals.totalTokens / aggregates.messages.total)
     : 0;
   const avgCost = aggregates.messages.total ? totals.totalCost / aggregates.messages.total : 0;
-  const cacheBase = totals.input + totals.cacheRead;
+  const cacheBase = totals.input + totals.cacheRead + totals.cacheWrite;
   const cacheHitRate = cacheBase > 0 ? totals.cacheRead / cacheBase : 0;
   const cacheHitLabel =
     cacheBase > 0 ? `${(cacheHitRate * 100).toFixed(1)}%` : t("usage.common.emptyValue");
