diff --git a/apps/web/src/pages/insights-page-render.test.tsx b/apps/web/src/pages/insights-page-render.test.tsx
index d7e1c994..4be3a9fa 100644
--- a/apps/web/src/pages/insights-page-render.test.tsx
+++ b/apps/web/src/pages/insights-page-render.test.tsx
@@ -134,4 +134,61 @@ describe("InsightsPage 'Running now' strip", () => {
     expect(el.textContent).toContain("1 in progress");
     expect(el.textContent).toContain("Weekly digest");
   });
+
+  // Liveness is not a windowed property: a run that started long before the
+  // 7-day window and is still running must not disappear from the strip or
+  // read 0 in the "Running now" KPI just because its start time falls
+  // outside `range`.
+  test("a run started 8 days ago that is still running stays in the strip and the KPI", () => {
+    const eightDaysAgo = new Date(
+      Date.now() - 8 * 24 * 60 * 60 * 1000,
+    ).toISOString();
+    const el = render("/insights", {
+      data: [
+        {
+          id: "run_long_haul",
+          tenantId: "tnt_bench_a",
+          definitionId: "wfd_a",
+          definitionName: "Long haul",
+          address: "addr",
+          status: "running",
+          createdAt: eightDaysAgo,
+          updatedAt: eightDaysAgo,
+          routineId: null,
+          routineName: null,
+        },
+      ],
+      nextCursor: null,
+    });
+    expect(el.textContent).toContain("Running now");
+    expect(el.textContent).toContain("1 in progress");
+    expect(el.textContent).toContain("Long haul");
+    expect(el.textContent).toContain("in flight");
+  });
+
+  test("the elapsed label ticks forward while a run is live, not frozen at first render", async () => {
+    const startedAt = new Date(Date.now() - 2_000).toISOString();
+    const el = render("/insights", {
+      data: [
+        {
+          id: "run_ticking",
+          tenantId: "tnt_bench_a",
+          definitionId: "wfd_a",
+          definitionName: "Weekly digest",
+          address: "addr",
+          status: "running",
+          createdAt: startedAt,
+          updatedAt: startedAt,
+          routineId: null,
+          routineName: null,
+        },
+      ],
+      nextCursor: null,
+    });
+    const before = el.textContent;
+    await act(async () => {
+      await new Promise((resolve) => setTimeout(resolve, 1_200));
+    });
+    expect(el.textContent).not.toBe(before);
+  });
 });
diff --git a/apps/web/src/pages/insights-page.tsx b/apps/web/src/pages/insights-page.tsx
index b04e84be..93a9485a 100644
--- a/apps/web/src/pages/insights-page.tsx
+++ b/apps/web/src/pages/insights-page.tsx
@@ -442,6 +442,29 @@ export function elapsedLabel(createdAt: string, now: number): string {
   return durationLabel(Math.max(0, now - startMs));
 }
 
+const ELAPSED_TICK_MS = 1_000;
+
+/** Ticks once a second while `enabled` — the clock the elapsed label next to
+ * the pulsing `StatusDot` reads from, so it counts up like the live indicator
+ * beside it instead of freezing at whatever instant this component mounted
+ * or last re-rendered for an unrelated reason. */
+function useTickingNow(enabled: boolean): number {
+  const [now, setNow] = useState(() => Date.now());
+  useEffect(() => {
+    if (!enabled) return undefined;
+    const timer = setInterval(() => setNow(Date.now()), ELAPSED_TICK_MS);
+    return () => clearInterval(timer);
+  }, [enabled]);
+  return now;
+}
+
+/** A run actually in flight right now (`status: running | updating`) —
+ * liveness is not a windowed property, so this filters the full run set,
+ * never the range-filtered one. */
+function isRunningNow(run: InsightsRun): boolean {
+  return run.status === "running" || run.status === "updating";
+}
+
 /**
  * "Running now" — a horizontally scrolling strip of the runs actually in
  * flight this instant (`status: running | updating`), not a fabricated
@@ -457,11 +480,9 @@ function RunningNowStrip({
   readonly runs: readonly InsightsRun[];
   readonly onOpenRun: (id: string) => void;
 }) {
-  const running = runs.filter(
-    (r) => r.status === "running" || r.status === "updating",
-  );
+  const running = runs.filter(isRunningNow);
+  const now = useTickingNow(running.length > 0);
   if (running.length === 0) return null;
-  const now = Date.now();
 
   return (
     <section className="insights-running-now" aria-label="Running now">
@@ -619,6 +640,7 @@ function InsightsLanding({
   byModel,
   byTool,
   runs,
+  runsNextCursor,
   routines,
   workbenches,
   latency,
@@ -633,6 +655,12 @@ function InsightsLanding({
   readonly byModel: readonly ModelUsage[] | null;
   readonly byTool: readonly ToolCall[] | null;
   readonly runs: readonly InsightsRun[];
+  /** The feed's own `nextCursor` (`limit=100` fetch, see
+   * `insightsTopLevelRunsPath`) — non-null means more runs exist in this
+   * window than the 100 fetched, so the KPIs/sparkline/outcome chart below
+   * disclose the cap instead of silently presenting a truncated series as
+   * complete. */
+  readonly runsNextCursor: string | null;
   readonly routines: readonly Routine[];
   /** Null while `/workbenches` hasn't resolved (or this landing is already
    * scoped to one workbench, where a breakdown of one has nothing to
@@ -653,6 +681,12 @@ function InsightsLanding({
   const stats = computeInsightsStats(windowedRuns, routines);
   const purposeRuns = purposeRunsForInsights(windowedRuns);
 
+  // Liveness is not a windowed property: a run that started before
+  // `range.from` and is still going is running right now regardless of when
+  // it started, so "Running now" reads off every fetched run, never the
+  // range-filtered subset above.
+  const runningNow = purposeRunsForInsights(runs).filter(isRunningNow);
+
   // Absent usage → zeros at the client boundary (never demo peaks / em-dash
   // for "no spend"). Real fetched summary is preserved when present.
   const usage = summary ?? EMPTY_OVERALL_USAGE;
@@ -726,17 +760,17 @@ function InsightsLanding({
             loading={loading}
           />
         ) : null}
-        {stats.running > 0 || loading ? (
+        {runningNow.length > 0 || loading ? (
           <InsightsStat
             label="Running now"
-            value={tileValue(formatCount(stats.running), loading)}
+            value={tileValue(formatCount(runningNow.length), loading)}
             detail="in flight"
             loading={loading}
           />
         ) : null}
       </StatGrid>
 
-      <RunningNowStrip runs={purposeRuns} onOpenRun={onOpenRun} />
+      <RunningNowStrip runs={runningNow} onOpenRun={onOpenRun} />
 
       {latency !== null && latency.total.samples > 0 ? (
         <StatGrid columns={4}>
@@ -780,6 +814,21 @@ function InsightsLanding({
         </p>
       ) : null}
 
+      {runsNextCursor !== null ? (
+        <p className="insights-note">
+          Runs, sparkline, and outcomes below reflect the 100 most recent runs —
+          more exist in this window.{" "}
+          <button
+            type="button"
+            className="font-semibold text-primary-emphasis"
+            onClick={onOpenRuns}
+          >
+            See all runs & traces
+          </button>
+          .
+        </p>
+      ) : null}
+
       <div className="insights-grid">
         <section className="insights-panel">
           <ActivityBars days={activityDays} />
@@ -1476,6 +1525,7 @@ export function InsightsPage({
             byModel={byModelData}
             byTool={byToolData}
             runs={runsData}
+            runsNextCursor={runsNextCursor}
             routines={routinesData}
             workbenches={workbenchesData}
             latency={latencyData}
diff --git a/apps/web/test/insights-page.test.tsx b/apps/web/test/insights-page.test.tsx
index fd8200b4..6edbe1f0 100644
--- a/apps/web/test/insights-page.test.tsx
+++ b/apps/web/test/insights-page.test.tsx
@@ -278,6 +278,51 @@ describe("InsightsPage scope switcher", () => {
   });
 });
 
+function renderAtPathWithRuns(path: string, nextCursor: string | null): string {
+  return renderToStaticMarkup(
+    <TestQueryProvider>
+      <NavigationProvider navigate={() => undefined}>
+        <BenchProvider>
+          <InsightsPage
+            path={path}
+            summary={{ kind: "ready", data: EMPTY_OVERALL_USAGE }}
+            activity={{ kind: "ready", data: [] }}
+            byTool={{ kind: "ready", data: [] }}
+            runs={{
+              kind: "ready",
+              data: { data: [purposeRun], nextCursor },
+            }}
+            routines={emptyRoutines}
+            workbenches={emptyWorkbenches}
+            latency={emptyLatency}
+            range={range}
+            scope={null}
+            resolveWorkbenchIdForTenant={() => null}
+            scopeLabel="All workbenches"
+          />
+        </BenchProvider>
+      </NavigationProvider>
+    </TestQueryProvider>,
+  );
+}
+
+describe("InsightsPage landing run-cap disclosure", () => {
+  // The feed is fetched at a fixed limit=100 (see insightsTopLevelRunsPath).
+  // A non-null nextCursor means the window truly holds more runs than were
+  // fetched, so the KPIs/sparkline/outcome chart built from that truncated
+  // set must say so rather than presenting it as the complete series —
+  // InsightsRunsHistory already discloses its own cap; the landing must too.
+  test("a non-null nextCursor discloses the 100-run cap on the landing view", () => {
+    const markup = renderAtPathWithRuns("/insights", "cursor_2");
+    expect(markup).toContain("100 most recent runs");
+  });
+
+  test("a null nextCursor (fewer than 100 runs) shows no cap disclosure", () => {
+    const markup = renderAtPathWithRuns("/insights", null);
+    expect(markup).not.toContain("100 most recent runs");
+  });
+});
+
 describe("InsightsPage breadcrumbs", () => {
   test("runs history puts an Insights / Run history trail in the top bar", () => {
     const markup = renderAtPath("/insights/runs");