Chat improvements (#23195)

* Support token streaming stats

* Propogate streaming token stats to chat calls

* Show token stats for each image

* Add settings to handle token stats and other options

* i18n

* Use select

* Improve mobile layout and spacing

frigate/api/chat.py

@@ -1411,6 +1411,11 @@ When a user refers to a specific object they have seen or describe with identify
                             )
                             + b"\n"
                         )
+                    elif kind == "stats":
+                        yield (
+                            json.dumps({"type": "stats", **value}).encode("utf-8")
+                            + b"\n"
+                        )
                     elif kind == "message":
                         msg = value
                         if msg.get("finish_reason") == "error":

frigate/genai/azure-openai.py

@@ -10,6 +10,7 @@ from openai import AzureOpenAI
 
 from frigate.config import GenAIProviderEnum
 from frigate.genai import GenAIClient, register_genai_provider
+from frigate.genai.openai import _stats_from_openai_usage
 
 logger = logging.getLogger(__name__)
 
@@ -210,6 +211,7 @@ class OpenAIClient(GenAIClient):
                 "messages": messages,
                 "timeout": self.timeout,
                 "stream": True,
+                "stream_options": {"include_usage": True},
             }
 
             if tools:
@@ -221,10 +223,15 @@ class OpenAIClient(GenAIClient):
             content_parts: list[str] = []
             tool_calls_by_index: dict[int, dict[str, Any]] = {}
             finish_reason = "stop"
+            usage_stats: Optional[dict[str, Any]] = None
 
             stream = self.provider.chat.completions.create(**request_params)  # type: ignore[call-overload]
 
             for chunk in stream:
+                chunk_usage = getattr(chunk, "usage", None)
+                if chunk_usage is not None:
+                    usage_stats = _stats_from_openai_usage(chunk_usage)
+
                 if not chunk or not chunk.choices:
                     continue
 
@@ -284,6 +291,9 @@ class OpenAIClient(GenAIClient):
                     )
                 finish_reason = "tool_calls"
 
+            if usage_stats is not None:
+                yield ("stats", usage_stats)
+
             yield (
                 "message",
                 {

frigate/genai/gemini.py

@@ -14,6 +14,20 @@ from frigate.genai import GenAIClient, register_genai_provider
 logger = logging.getLogger(__name__)
 
 
+def _stats_from_gemini_usage(usage: Any) -> Optional[dict[str, Any]]:
+    """Build a stats dict from a Gemini usage_metadata object."""
+    prompt_tokens = getattr(usage, "prompt_token_count", None)
+    completion_tokens = getattr(usage, "candidates_token_count", None)
+    if prompt_tokens is None and completion_tokens is None:
+        return None
+    stats: dict[str, Any] = {}
+    if isinstance(prompt_tokens, int):
+        stats["prompt_tokens"] = prompt_tokens
+    if isinstance(completion_tokens, int):
+        stats["completion_tokens"] = completion_tokens
+    return stats or None
+
+
 @register_genai_provider(GenAIProviderEnum.gemini)
 class GeminiClient(GenAIClient):
     """Generative AI client for Frigate using Gemini."""
@@ -471,6 +485,7 @@ class GeminiClient(GenAIClient):
             content_parts: list[str] = []
             tool_calls_by_index: dict[int, dict[str, Any]] = {}
             finish_reason = "stop"
+            usage_stats: Optional[dict[str, Any]] = None
 
             stream = await self.provider.aio.models.generate_content_stream(
                 model=self.genai_config.model,
@@ -479,6 +494,12 @@ class GeminiClient(GenAIClient):
             )
 
             async for chunk in stream:
+                chunk_usage = getattr(chunk, "usage_metadata", None)
+                if chunk_usage is not None:
+                    maybe_stats = _stats_from_gemini_usage(chunk_usage)
+                    if maybe_stats is not None:
+                        usage_stats = maybe_stats
+
                 if not chunk or not chunk.candidates:
                     continue
 
@@ -565,6 +586,9 @@ class GeminiClient(GenAIClient):
                     )
                 finish_reason = "tool_calls"
 
+            if usage_stats is not None:
+                yield ("stats", usage_stats)
+
             yield (
                 "message",
                 {

frigate/genai/llama_cpp.py

@@ -18,6 +18,52 @@ from frigate.genai.utils import parse_tool_calls_from_message
 logger = logging.getLogger(__name__)
 
 
+def _stats_from_llama_cpp_chunk(data: dict[str, Any]) -> Optional[dict[str, Any]]:
+    """Build a stats dict from a llama.cpp streaming chunk.
+
+    Final-chunk `usage` carries authoritative token counts. Per-chunk
+    `timings` (enabled via timings_per_token) carries the running token
+    counts (prompt_n, predicted_n) and generation rate, so live updates
+    work mid-stream.
+    """
+    usage = data.get("usage") or {}
+    timings = data.get("timings") or {}
+    prompt_tokens = usage.get("prompt_tokens")
+    completion_tokens = usage.get("completion_tokens")
+    predicted_ms = timings.get("predicted_ms")
+    tps = timings.get("predicted_per_second")
+    stats: dict[str, Any] = {}
+
+    if not isinstance(prompt_tokens, int):
+        prompt_n = timings.get("prompt_n")
+
+        if isinstance(prompt_n, int):
+            prompt_tokens = prompt_n
+
+    if not isinstance(completion_tokens, int):
+        predicted_n = timings.get("predicted_n")
+
+        if isinstance(predicted_n, int):
+            completion_tokens = predicted_n
+
+    if not isinstance(prompt_tokens, int) and not isinstance(completion_tokens, int):
+        return None
+
+    if isinstance(prompt_tokens, int):
+        stats["prompt_tokens"] = prompt_tokens
+
+    if isinstance(completion_tokens, int):
+        stats["completion_tokens"] = completion_tokens
+
+    if isinstance(predicted_ms, (int, float)) and predicted_ms > 0:
+        stats["completion_duration_ms"] = float(predicted_ms)
+
+    if isinstance(tps, (int, float)) and tps > 0:
+        stats["tokens_per_second"] = float(tps)
+
+    return stats or None
+
+
 def _parse_launch_arg(args: list[str], flag: str) -> str | None:
     """Return the value following `flag` in a positional argv list, or None."""
     try:
@@ -462,6 +508,8 @@ class LlamaCppClient(GenAIClient):
         }
         if stream:
             payload["stream"] = True
+            payload["stream_options"] = {"include_usage": True}
+            payload["timings_per_token"] = True
         if tools:
             payload["tools"] = tools
             if openai_tool_choice is not None:
@@ -724,6 +772,9 @@ class LlamaCppClient(GenAIClient):
                             data = json.loads(data_str)
                         except json.JSONDecodeError:
                             continue
+                        maybe_stats = _stats_from_llama_cpp_chunk(data)
+                        if maybe_stats is not None:
+                            yield ("stats", maybe_stats)
                         choices = data.get("choices") or []
                         if not choices:
                             continue

frigate/genai/ollama.py

@@ -18,6 +18,37 @@ from frigate.genai.utils import parse_tool_calls_from_message
 logger = logging.getLogger(__name__)
 
 
+def _extract_ollama_stats(response: Any) -> Optional[dict[str, Any]]:
+    """Build a stats dict from Ollama's response metadata.
+
+    Ollama reports eval_count/eval_duration (generation) and
+    prompt_eval_count (context size). Durations are nanoseconds.
+    """
+    if not response:
+        return None
+    if hasattr(response, "get"):
+        getter = response.get
+    else:
+        getter = lambda key: getattr(response, key, None)  # noqa: E731
+
+    eval_count = getter("eval_count")
+    eval_duration_ns = getter("eval_duration")
+    prompt_eval_count = getter("prompt_eval_count")
+    if eval_count is None and prompt_eval_count is None:
+        return None
+
+    stats: dict[str, Any] = {}
+    if isinstance(prompt_eval_count, int):
+        stats["prompt_tokens"] = prompt_eval_count
+    if isinstance(eval_count, int):
+        stats["completion_tokens"] = eval_count
+    if isinstance(eval_duration_ns, int) and eval_duration_ns > 0:
+        stats["completion_duration_ms"] = eval_duration_ns / 1_000_000
+        if isinstance(eval_count, int) and eval_count > 0:
+            stats["tokens_per_second"] = eval_count / (eval_duration_ns / 1_000_000_000)
+    return stats or None
+
+
 def _normalize_multimodal_content(
     content: Any,
 ) -> tuple[Optional[str], Optional[list[bytes]]]:
@@ -403,6 +434,9 @@ class OllamaClient(GenAIClient):
                 content = result.get("content")
                 if content:
                     yield ("content_delta", content)
+                stats = _extract_ollama_stats(response)
+                if stats is not None:
+                    yield ("stats", stats)
                 yield ("message", result)
                 return
 
@@ -416,6 +450,7 @@ class OllamaClient(GenAIClient):
             )
             content_parts: list[str] = []
             final_message: dict[str, Any] | None = None
+            final_chunk: Any = None
             stream = await async_client.chat(**request_params)
             async for chunk in stream:
                 if not chunk or "message" not in chunk:
@@ -426,6 +461,7 @@ class OllamaClient(GenAIClient):
                     content_parts.append(delta)
                     yield ("content_delta", delta)
                 if chunk.get("done"):
+                    final_chunk = chunk
                     full_content = "".join(content_parts).strip() or None
                     final_message = {
                         "content": full_content,
@@ -434,6 +470,10 @@ class OllamaClient(GenAIClient):
                     }
                     break
 
+            stats = _extract_ollama_stats(final_chunk)
+            if stats is not None:
+                yield ("stats", stats)
+
             if final_message is not None:
                 yield ("message", final_message)
             else:

frigate/genai/openai.py

@@ -14,6 +14,22 @@ from frigate.genai import GenAIClient, register_genai_provider
 logger = logging.getLogger(__name__)
 
 
+def _stats_from_openai_usage(usage: Any) -> Optional[dict[str, Any]]:
+    """Build a stats dict from an OpenAI-compatible usage object."""
+    if usage is None:
+        return None
+    prompt_tokens = getattr(usage, "prompt_tokens", None)
+    completion_tokens = getattr(usage, "completion_tokens", None)
+    if prompt_tokens is None and completion_tokens is None:
+        return None
+    stats: dict[str, Any] = {}
+    if isinstance(prompt_tokens, int):
+        stats["prompt_tokens"] = prompt_tokens
+    if isinstance(completion_tokens, int):
+        stats["completion_tokens"] = completion_tokens
+    return stats or None
+
+
 @register_genai_provider(GenAIProviderEnum.openai)
 class OpenAIClient(GenAIClient):
     """Generative AI client for Frigate using OpenAI."""
@@ -298,6 +314,7 @@ class OpenAIClient(GenAIClient):
                 "messages": messages,
                 "timeout": self.timeout,
                 "stream": True,
+                "stream_options": {"include_usage": True},
             }
 
             if tools:
@@ -318,10 +335,15 @@ class OpenAIClient(GenAIClient):
             content_parts: list[str] = []
             tool_calls_by_index: dict[int, dict[str, Any]] = {}
             finish_reason = "stop"
+            usage_stats: Optional[dict[str, Any]] = None
 
             stream = self.provider.chat.completions.create(**request_params)  # type: ignore[call-overload]
 
             for chunk in stream:
+                chunk_usage = getattr(chunk, "usage", None)
+                if chunk_usage is not None:
+                    usage_stats = _stats_from_openai_usage(chunk_usage)
+
                 if not chunk or not chunk.choices:
                     continue
 
@@ -381,6 +403,9 @@ class OpenAIClient(GenAIClient):
                     )
                 finish_reason = "tool_calls"
 
+            if usage_stats is not None:
+                yield ("stats", usage_stats)
+
             yield (
                 "message",
                 {

web/public/locales/en/views/chat.json

@@ -42,5 +42,23 @@
     "show_camera_status": "What is the current status of my cameras?",
     "recap": "What happened while I was away?",
     "watch_camera": "Watch the front door and let me know if anyone shows up"
+  },
+  "new_chat": "New chat",
+  "settings": {
+    "title": "Chat settings",
+    "show_stats": {
+      "title": "Show stats",
+      "desc": "Show generation rate and context size for chat responses.",
+      "while_generating": "While generating",
+      "always": "Always"
+    },
+    "auto_scroll": {
+      "title": "Auto-scroll",
+      "desc": "Follow new messages as they arrive."
+    }
+  },
+  "stats": {
+    "context": "{{tokens}} tokens",
+    "tokens_per_second": "{{rate}} t/s"
   }
 }

web/src/components/chat/ChatEventThumbnailsRow.tsx

@@ -6,7 +6,6 @@ import {
   TooltipContent,
   TooltipTrigger,
 } from "@/components/ui/tooltip";
-import { cn } from "@/lib/utils";
 
 type ChatEvent = { id: string; score?: number };
 
@@ -37,10 +36,7 @@ export function ChatEventThumbnailsRow({
   const renderThumb = (event: ChatEvent, isAnchor = false) => (
     <div
       key={event.id}
-      className={cn(
+      className="relative aspect-square size-32 shrink-0 overflow-hidden rounded-lg"
-        "relative aspect-square size-32 shrink-0 overflow-hidden rounded-lg",
-        isAnchor && "ring-2 ring-primary",
-      )}
     >
       <button
         type="button"
@@ -71,9 +67,15 @@ export function ChatEventThumbnailsRow({
         <TooltipContent>{t("open_in_explore")}</TooltipContent>
       </Tooltip>
       {isAnchor && (
+        <>
+          <span
+            aria-hidden="true"
+            className="pointer-events-none absolute inset-0 rounded-lg ring-2 ring-inset ring-primary"
+          />
           <span className="pointer-events-none absolute left-1 top-1 rounded bg-primary px-1 text-[10px] text-primary-foreground">
             {t("anchor")}
           </span>
+        </>
       )}
     </div>
   );

web/src/components/chat/ChatMessage.tsx

@@ -17,6 +17,7 @@ import {
 import { cn } from "@/lib/utils";
 import { ChatAttachmentChip } from "@/components/chat/ChatAttachmentChip";
 import { parseAttachedEvent } from "@/utils/chatUtil";
+import type { ChatStats, ShowStatsMode } from "@/types/chat";
 
 type MessageBubbleProps = {
   role: "user" | "assistant";
@@ -24,14 +25,29 @@ type MessageBubbleProps = {
   messageIndex?: number;
   onEditSubmit?: (messageIndex: number, newContent: string) => void;
   isComplete?: boolean;
+  stats?: ChatStats;
+  showStats?: ShowStatsMode;
 };
 
+function formatTokens(n: number | undefined): string | null {
+  if (n === undefined) return null;
+  if (n >= 1000) return `${(n / 1000).toFixed(1)}k`;
+  return String(n);
+}
+
+function formatRate(rate: number | undefined): string | null {
+  if (rate === undefined || rate <= 0) return null;
+  return rate >= 10 ? rate.toFixed(0) : rate.toFixed(1);
+}
+
 export function MessageBubble({
   role,
   content,
   messageIndex = 0,
   onEditSubmit,
   isComplete = true,
+  stats,
+  showStats = "while_generating",
 }: MessageBubbleProps) {
   const { t } = useTranslation(["views/chat", "common"]);
   const isUser = role === "user";
@@ -214,7 +230,7 @@ export function MessageBubble({
           </div>
         )}
       </div>
-      <div className="flex items-center gap-0.5">
+      <div className="flex items-center gap-1.5">
         {isUser && onEditSubmit != null && (
           <Tooltip>
             <TooltipTrigger asChild>
@@ -256,6 +272,27 @@ export function MessageBubble({
             </TooltipContent>
           </Tooltip>
         )}
+        {!isUser &&
+          stats &&
+          (showStats === "always" || !isComplete) &&
+          (() => {
+            const ctx = formatTokens(stats.promptTokens);
+            const rate = formatRate(stats.tokensPerSecond);
+            if (ctx === null && rate === null) return null;
+            return (
+              <div className="flex items-center gap-1 text-xs text-muted-foreground">
+                {ctx !== null && (
+                  <span>{t("stats.context", { tokens: ctx })}</span>
+                )}
+                {ctx !== null && rate !== null && (
+                  <span aria-hidden="true">·</span>
+                )}
+                {rate !== null && (
+                  <span>{t("stats.tokens_per_second", { rate })}</span>
+                )}
+              </div>
+            );
+          })()}
       </div>
     </div>
   );

web/src/components/chat/ChatSettings.tsx

@@ -0,0 +1,108 @@
+import { Button } from "@/components/ui/button";
+import { useState } from "react";
+import { isDesktop } from "react-device-detect";
+import { cn } from "@/lib/utils";
+import PlatformAwareDialog from "../overlay/dialog/PlatformAwareDialog";
+import { FaCog } from "react-icons/fa";
+import {
+  Select,
+  SelectContent,
+  SelectGroup,
+  SelectItem,
+  SelectTrigger,
+} from "@/components/ui/select";
+import { Switch } from "@/components/ui/switch";
+import { Label } from "@/components/ui/label";
+import { DropdownMenuSeparator } from "@/components/ui/dropdown-menu";
+import { useTranslation } from "react-i18next";
+import type { ShowStatsMode } from "@/types/chat";
+
+type ChatSettingsProps = {
+  showStats: ShowStatsMode;
+  setShowStats: (mode: ShowStatsMode) => void;
+  autoScroll: boolean;
+  setAutoScroll: (enabled: boolean) => void;
+};
+
+export default function ChatSettings({
+  showStats,
+  setShowStats,
+  autoScroll,
+  setAutoScroll,
+}: ChatSettingsProps) {
+  const { t } = useTranslation(["views/chat"]);
+  const [open, setOpen] = useState(false);
+
+  const trigger = (
+    <Button
+      className="flex items-center md:gap-2"
+      aria-label={t("settings.title")}
+      size="sm"
+    >
+      <FaCog className="text-secondary-foreground" />
+      <span className="hidden md:inline">{t("settings.title")}</span>
+    </Button>
+  );
+
+  const content = (
+    <div className="my-3 space-y-5 py-3 md:mt-0 md:py-0">
+      <div className="space-y-3">
+        <div className="space-y-0.5">
+          <div className="text-md">{t("settings.show_stats.title")}</div>
+          <div className="text-xs text-muted-foreground">
+            {t("settings.show_stats.desc")}
+          </div>
+        </div>
+        <Select
+          value={showStats}
+          onValueChange={(v) => setShowStats(v as ShowStatsMode)}
+        >
+          <SelectTrigger className="w-full">
+            {showStats === "always"
+              ? t("settings.show_stats.always")
+              : t("settings.show_stats.while_generating")}
+          </SelectTrigger>
+          <SelectContent>
+            <SelectGroup>
+              <SelectItem className="cursor-pointer" value="while_generating">
+                {t("settings.show_stats.while_generating")}
+              </SelectItem>
+              <SelectItem className="cursor-pointer" value="always">
+                {t("settings.show_stats.always")}
+              </SelectItem>
+            </SelectGroup>
+          </SelectContent>
+        </Select>
+      </div>
+      <DropdownMenuSeparator />
+      <div className="flex items-center justify-between gap-3">
+        <div className="space-y-0.5">
+          <Label htmlFor="auto-scroll" className="text-md cursor-pointer">
+            {t("settings.auto_scroll.title")}
+          </Label>
+          <div className="text-xs text-muted-foreground">
+            {t("settings.auto_scroll.desc")}
+          </div>
+        </div>
+        <Switch
+          id="auto-scroll"
+          checked={autoScroll}
+          onCheckedChange={setAutoScroll}
+        />
+      </div>
+    </div>
+  );
+
+  return (
+    <PlatformAwareDialog
+      trigger={trigger}
+      content={content}
+      contentClassName={cn(
+        "scrollbar-container h-auto overflow-y-auto",
+        isDesktop ? "max-h-[80dvh] w-72" : "px-4",
+      )}
+      open={open}
+      onOpenChange={setOpen}
+    />
+  );
+}

web/src/components/chat/ChatStartingState.tsx

@@ -54,7 +54,9 @@ export function ChatStartingState({ onSendMessage }: ChatStartingStateProps) {
     <div className="flex size-full flex-col items-center justify-center gap-6 p-8">
       <div className="flex flex-col items-center gap-2">
         <h1 className="text-4xl font-bold text-foreground">{t("title")}</h1>
-        <p className="text-muted-foreground">{t("subtitle")}</p>
+        <p className="text-center text-muted-foreground md:text-left">
+          {t("subtitle")}
+        </p>
       </div>
 
       <div className="flex w-full max-w-2xl flex-col items-center gap-4">

web/src/pages/Chat.tsx

@@ -1,7 +1,7 @@
 import { Button } from "@/components/ui/button";
 import { Input } from "@/components/ui/input";
 import { FaArrowUpLong, FaStop } from "react-icons/fa6";
-import { LuCircleAlert } from "react-icons/lu";
+import { LuCircleAlert, LuMessageSquarePlus } from "react-icons/lu";
 import { useTranslation } from "react-i18next";
 import { useState, useCallback, useRef, useEffect, useMemo } from "react";
 import axios from "axios";
@@ -12,7 +12,9 @@ import { ChatStartingState } from "@/components/chat/ChatStartingState";
 import { ChatAttachmentChip } from "@/components/chat/ChatAttachmentChip";
 import { ChatQuickReplies } from "@/components/chat/ChatQuickReplies";
 import { ChatPaperclipButton } from "@/components/chat/ChatPaperclipButton";
-import type { ChatMessage } from "@/types/chat";
+import ChatSettings from "@/components/chat/ChatSettings";
+import type { ChatMessage, ShowStatsMode } from "@/types/chat";
+import { usePersistence } from "@/hooks/use-persistence";
 import {
   getEventIdsFromSearchObjectsToolCalls,
   getFindSimilarObjectsFromToolCalls,
@@ -27,6 +29,14 @@ export default function ChatPage() {
   const [isLoading, setIsLoading] = useState(false);
   const [error, setError] = useState<string | null>(null);
   const [attachedEventId, setAttachedEventId] = useState<string | null>(null);
+  const [showStats, setShowStats] = usePersistence<ShowStatsMode>(
+    "chat-show-stats",
+    "while_generating",
+  );
+  const [autoScroll, setAutoScroll] = usePersistence<boolean>(
+    "chat-auto-scroll",
+    true,
+  );
   const scrollRef = useRef<HTMLDivElement>(null);
   const abortRef = useRef<AbortController | null>(null);
 
@@ -36,13 +46,14 @@ export default function ChatPage() {
 
   // Auto-scroll to bottom when messages change, but only if near bottom
   useEffect(() => {
+    if (!autoScroll) return;
     const el = scrollRef.current;
     if (!el) return;
     const isNearBottom = el.scrollHeight - el.scrollTop - el.clientHeight < 150;
     if (isNearBottom) {
       el.scrollTo({ top: el.scrollHeight, behavior: "smooth" });
     }
-  }, [messages]);
+  }, [messages, autoScroll]);
 
   const submitConversation = useCallback(
     async (messagesToSend: ChatMessage[]) => {
@@ -125,6 +136,16 @@ export default function ChatPage() {
     setIsLoading(false);
   }, []);
 
+  const startNewChat = useCallback(() => {
+    abortRef.current?.abort();
+    abortRef.current = null;
+    setIsLoading(false);
+    setMessages([]);
+    setInput("");
+    setAttachedEventId(null);
+    setError(null);
+  }, []);
+
   const handleEditSubmit = useCallback(
     (messageIndex: number, newContent: string) => {
       const newList: ChatMessage[] = [
@@ -140,28 +161,44 @@ export default function ChatPage() {
     setAttachedEventId(null);
   }, []);
 
+  const hasStarted = messages.length > 0;
+
   return (
-    <div className="flex size-full justify-center p-2 md:p-4">
+    <div className="flex size-full flex-col">
-      <div className="flex size-full flex-col xl:w-[50%] 3xl:w-[35%]">
+      <div className="flex shrink-0 items-center justify-end gap-2 px-2 pb-3 pt-2 md:px-4 md:pt-4">
-        {messages.length === 0 ? (
+        {hasStarted && (
-          <ChatStartingState
+          <Button
-            onSendMessage={(message) => {
+            className="flex items-center md:gap-2"
-              setInput("");
+            aria-label={t("new_chat")}
-              submitConversation([{ role: "user", content: message }]);
+            size="sm"
-            }}
+            onClick={startNewChat}
+          >
+            <LuMessageSquarePlus className="text-secondary-foreground" />
+            <span className="hidden md:inline">{t("new_chat")}</span>
+          </Button>
+        )}
+        <ChatSettings
+          showStats={showStats ?? "while_generating"}
+          setShowStats={setShowStats}
+          autoScroll={autoScroll ?? true}
+          setAutoScroll={setAutoScroll}
         />
-        ) : (
+      </div>
-          <>
       <div
         ref={scrollRef}
-              className="scrollbar-container flex min-h-0 w-full flex-1 flex-col gap-3 overflow-y-auto"
+        className="scrollbar-container flex min-h-0 flex-1 flex-col overflow-y-auto"
       >
+        <div className="flex flex-1 justify-center px-2 md:px-4">
+          <div className="flex w-full flex-col xl:w-[50%] 3xl:w-[35%]">
+            {hasStarted ? (
+              <div className="flex w-full flex-1 flex-col gap-3 pb-3">
                 {messages.map((msg, i) => {
                   const isLastAssistant =
                     i === messages.length - 1 && msg.role === "assistant";
                   const isComplete =
                     msg.role === "user" || !isLoading || !isLastAssistant;
-                const hasToolCalls = msg.toolCalls && msg.toolCalls.length > 0;
+                  const hasToolCalls =
+                    msg.toolCalls && msg.toolCalls.length > 0;
                   const hasContent = !!msg.content?.trim();
                   const showProcessing =
                     isLastAssistant && isLoading && !hasContent;
@@ -204,6 +241,8 @@ export default function ChatPage() {
                             msg.role === "user" ? handleEditSubmit : undefined
                           }
                           isComplete={isComplete}
+                          stats={msg.stats}
+                          showStats={showStats}
                         />
                       )}
                       {msg.role === "assistant" &&
@@ -244,9 +283,20 @@ export default function ChatPage() {
                   </p>
                 )}
               </div>
-          </>
+            ) : (
+              <ChatStartingState
+                onSendMessage={(message) => {
+                  setInput("");
+                  submitConversation([{ role: "user", content: message }]);
+                }}
+              />
             )}
-        {messages.length > 0 && (
+          </div>
+        </div>
+      </div>
+      {hasStarted && (
+        <div className="flex shrink-0 justify-center p-2 md:px-4 md:pb-4">
+          <div className="flex w-full xl:w-[50%] 3xl:w-[35%]">
             <ChatEntry
               input={input}
               setInput={setInput}
@@ -259,9 +309,10 @@ export default function ChatPage() {
               onStop={stopGeneration}
               recentEventIds={recentEventIds}
             />
-        )}
           </div>
         </div>
+      )}
+    </div>
   );
 }
 
@@ -298,7 +349,7 @@ function ChatEntry({
   };
 
   return (
-    <div className="mt-2 flex w-full flex-col items-stretch justify-center gap-2 rounded-xl bg-secondary p-3">
+    <div className="flex w-full flex-col items-stretch justify-center gap-2 rounded-xl bg-secondary p-3">
       {attachedEventId && (
         <div className="flex items-center">
           <ChatAttachmentChip

web/src/types/chat.ts

@@ -8,9 +8,19 @@ export type ChatMessage = {
   role: "user" | "assistant";
   content: string;
   toolCalls?: ToolCall[];
+  stats?: ChatStats;
 };
 
 export type StartingRequest = {
   label: string;
   prompt: string;
 };
+
+export type ChatStats = {
+  promptTokens?: number;
+  completionTokens?: number;
+  completionDurationMs?: number;
+  tokensPerSecond?: number;
+};
+
+export type ShowStatsMode = "while_generating" | "always";

web/src/utils/chatUtil.ts

@@ -1,4 +1,4 @@
-import type { ChatMessage, ToolCall } from "@/types/chat";
+import type { ChatMessage, ChatStats, ToolCall } from "@/types/chat";
 
 export type StreamChatCallbacks = {
   /** Update the messages array (e.g. pass to setState). */
@@ -7,14 +7,27 @@ export type StreamChatCallbacks = {
   onError: (message: string) => void;
   /** Called when the stream finishes (success or error). */
   onDone: () => void;
+  /** Called when the stream emits token/timing stats. The stats are also
+   * attached to the last assistant message in updateMessages, so consumers
+   * can usually rely on the message itself rather than wiring this up. */
+  onStats?: (stats: ChatStats) => void;
   /** Message used when fetch throws and no server error is available. */
   defaultErrorMessage?: string;
 };
 
+type StatsChunk = {
+  type: "stats";
+  prompt_tokens?: number;
+  completion_tokens?: number;
+  completion_duration_ms?: number;
+  tokens_per_second?: number;
+};
+
 type StreamChunk =
   | { type: "error"; error: string }
   | { type: "tool_calls"; tool_calls: ToolCall[] }
-  | { type: "content"; delta: string };
+  | { type: "content"; delta: string }
+  | StatsChunk;
 
 /**
  * POST to chat/completion with stream: true, parse NDJSON stream, and invoke
@@ -31,6 +44,7 @@ export async function streamChatCompletion(
     updateMessages,
     onError,
     onDone,
+    onStats,
     defaultErrorMessage = "Something went wrong. Please try again.",
   } = callbacks;
 
@@ -95,6 +109,23 @@ export async function streamChatCompletion(
         });
         return "continue";
       }
+      if (data.type === "stats") {
+        const stats: ChatStats = {
+          promptTokens: data.prompt_tokens,
+          completionTokens: data.completion_tokens,
+          completionDurationMs: data.completion_duration_ms,
+          tokensPerSecond: data.tokens_per_second,
+        };
+        updateMessages((prev) => {
+          const next = [...prev];
+          const lastMsg = next[next.length - 1];
+          if (lastMsg?.role === "assistant")
+            next[next.length - 1] = { ...lastMsg, stats };
+          return next;
+        });
+        onStats?.(stats);
+        return "continue";
+      }
       return "continue";
     };