Add streaming responses, switch to Haiku, user feedback, and startup data gathering

- Switch LLM from Claude Sonnet to Claude Haiku 3.5 (claude-haiku-4-5-20251001),
  reducing avg eval latency from ~7s to 3.8s with 100% pass rate (55/55)
- Add streaming endpoint (POST /ai/agent/stream) using Vercel AI SDK streamText()
  with SSE, so tokens render progressively in the chat UI
- Update Angular chat component to consume SSE via fetch + ReadableStream reader
- Add user feedback (thumbs up/down) buttons on assistant messages that POST to
  new /ai/feedback endpoint and log scores to Langfuse via REST API
- Add delayed loadCurrencies() call on startup to ensure exchange rates are
  populated after data-gathering queue processes on fresh deployments
- Fix eval suite dotenv loading so LLM-as-judge can read ANTHROPIC_API_KEY

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

apps/api/src/app/endpoints/ai/ai.controller.ts

@@ -13,10 +13,12 @@ import {
   Param,
   Post,
   Query,
+  Res,
   UseGuards
 } from '@nestjs/common';
 import { REQUEST } from '@nestjs/core';
 import { AuthGuard } from '@nestjs/passport';
+import { Response } from 'express';
 
 import { AiService } from './ai.service';
 
@@ -73,4 +75,49 @@ export class AiController {
       userId: this.request.user.id
     });
   }
+
+  @Post('agent/stream')
+  @HasPermission(permissions.readAiPrompt)
+  @UseGuards(AuthGuard('jwt'), HasPermissionGuard)
+  public async agentChatStream(
+    @Body() body: { message: string; conversationHistory?: any[] },
+    @Res() res: Response
+  ) {
+    res.setHeader('Content-Type', 'text/event-stream');
+    res.setHeader('Cache-Control', 'no-cache');
+    res.setHeader('Connection', 'keep-alive');
+    res.flushHeaders();
+
+    await this.aiService.agentChatStream({
+      message: body.message,
+      conversationHistory: body.conversationHistory,
+      impersonationId: undefined,
+      userCurrency: this.request.user.settings.settings.baseCurrency,
+      userId: this.request.user.id,
+      onChunk: (text) => {
+        res.write(`event: text\ndata: ${JSON.stringify(text)}\n\n`);
+      },
+      onDone: (metadata) => {
+        res.write(`event: done\ndata: ${JSON.stringify(metadata)}\n\n`);
+        res.end();
+      },
+      onError: (error) => {
+        res.write(`event: error\ndata: ${JSON.stringify({ error })}\n\n`);
+        res.end();
+      }
+    });
+  }
+
+  @Post('feedback')
+  @HasPermission(permissions.readAiPrompt)
+  @UseGuards(AuthGuard('jwt'), HasPermissionGuard)
+  public async submitFeedback(
+    @Body() body: { traceId: string; value: number }
+  ) {
+    return this.aiService.submitFeedback({
+      traceId: body.traceId,
+      value: body.value,
+      userId: this.request.user.id
+    });
+  }
 }

apps/api/src/app/endpoints/ai/ai.service.ts

@@ -11,20 +11,21 @@ import {
 import { Filter } from '@ghostfolio/common/interfaces';
 import type { AiPromptMode } from '@ghostfolio/common/types';
 
-import { Injectable, Logger } from '@nestjs/common';
 import { createAnthropic } from '@ai-sdk/anthropic';
+import { Injectable, Logger } from '@nestjs/common';
 import { createOpenRouter } from '@openrouter/ai-sdk-provider';
-import { generateText, CoreMessage } from 'ai';
+import { generateText, streamText, CoreMessage } from 'ai';
+import { randomUUID } from 'crypto';
 import type { ColumnDescriptor } from 'tablemark';
 
-import { getPortfolioHoldingsTool } from './tools/portfolio-holdings.tool';
-import { getPortfolioPerformanceTool } from './tools/portfolio-performance.tool';
 import { getAccountSummaryTool } from './tools/account-summary.tool';
 import { getDividendSummaryTool } from './tools/dividend-summary.tool';
-import { getTransactionHistoryTool } from './tools/transaction-history.tool';
-import { getLookupMarketDataTool } from './tools/market-data.tool';
 import { getExchangeRateTool } from './tools/exchange-rate.tool';
+import { getLookupMarketDataTool } from './tools/market-data.tool';
+import { getPortfolioHoldingsTool } from './tools/portfolio-holdings.tool';
+import { getPortfolioPerformanceTool } from './tools/portfolio-performance.tool';
 import { getPortfolioReportTool } from './tools/portfolio-report.tool';
+import { getTransactionHistoryTool } from './tools/transaction-history.tool';
 import { runVerificationChecks } from './verification';
 
 function getAgentSystemPrompt() {
@@ -197,36 +198,32 @@ export class AiService {
     return [
       `You are a neutral financial assistant. Please analyze the following investment portfolio (base currency being ${userCurrency}) in simple words.`,
       holdingsTableString,
-      "Structure your answer with these sections:",
+      'Structure your answer with these sections:',
-      "Overview: Briefly summarize the portfolio composition and allocation rationale.",
+      'Overview: Briefly summarize the portfolio composition and allocation rationale.',
-      "Risk Assessment: Identify potential risks, including market volatility, concentration, and sectoral imbalances.",
+      'Risk Assessment: Identify potential risks, including market volatility, concentration, and sectoral imbalances.',
-      "Advantages: Highlight strengths, focusing on growth potential, diversification, or other benefits.",
+      'Advantages: Highlight strengths, focusing on growth potential, diversification, or other benefits.',
-      "Disadvantages: Point out weaknesses, such as overexposure or lack of defensive assets.",
+      'Disadvantages: Point out weaknesses, such as overexposure or lack of defensive assets.',
-      "Target Group: Discuss who this portfolio might suit (e.g., risk tolerance, investment goals, life stages, and experience levels).",
+      'Target Group: Discuss who this portfolio might suit (e.g., risk tolerance, investment goals, life stages, and experience levels).',
-      "Optimization Ideas: Offer ideas to complement the portfolio, ensuring they are constructive and neutral in tone.",
+      'Optimization Ideas: Offer ideas to complement the portfolio, ensuring they are constructive and neutral in tone.',
-      "Conclusion: Provide a concise summary highlighting key insights.",
+      'Conclusion: Provide a concise summary highlighting key insights.',
       `Provide your answer in the following language: ${languageCode}.`
-    ].join("\n");
+    ].join('\n');
   }
 
-  public async agentChat({
+  private buildAgentConfig({
-    conversationHistory,
+    userId,
-    message,
     impersonationId,
-    userCurrency,
+    userCurrency
-    userId
   }: {
-    conversationHistory?: CoreMessage[];
+    userId: string;
-    message: string;
     impersonationId?: string;
     userCurrency: string;
-    userId: string;
   }) {
     const anthropicApiKey = process.env.ANTHROPIC_API_KEY;
 
     if (!anthropicApiKey) {
       throw new Error(
-        "ANTHROPIC_API_KEY is not configured. Please set the environment variable."
+        'ANTHROPIC_API_KEY is not configured. Please set the environment variable.'
       );
     }
 
@@ -272,23 +269,47 @@ export class AiService {
       })
     };
 
+    return { anthropic, tools };
+  }
+
+  public async agentChat({
+    conversationHistory,
+    message,
+    impersonationId,
+    userCurrency,
+    userId
+  }: {
+    conversationHistory?: CoreMessage[];
+    message: string;
+    impersonationId?: string;
+    userCurrency: string;
+    userId: string;
+  }) {
+    const { anthropic, tools } = this.buildAgentConfig({
+      userId,
+      impersonationId,
+      userCurrency
+    });
+
     const messages: CoreMessage[] = [
       ...(conversationHistory ?? []),
-      { role: "user" as const, content: message }
+      { role: 'user' as const, content: message }
     ];
 
+    const traceId = randomUUID();
+
     try {
       const result = await generateText({
-        model: anthropic("claude-sonnet-4-20250514"),
+        model: anthropic('claude-haiku-4-5-20251001'),
         system: getAgentSystemPrompt(),
         tools,
-        toolChoice: "auto",
+        toolChoice: 'auto',
         messages,
-        maxSteps: 5,
+        maxSteps: 10,
         experimental_telemetry: {
           isEnabled: true,
-          functionId: "ghostfolio-ai-agent",
+          functionId: 'ghostfolio-ai-agent',
-          metadata: { userId }
+          metadata: { userId, traceId }
         }
       });
 
@@ -299,12 +320,13 @@ export class AiService {
           args: tc.args
         }));
 
-      const toolResults = result.steps
+      const toolResults = result.steps.flatMap(
-        .flatMap((step) => step.toolResults ?? []);
+        (step) => step.toolResults ?? []
+      );
 
       const updatedHistory: CoreMessage[] = [
         ...messages,
-        { role: "assistant" as const, content: result.text }
+        { role: 'assistant' as const, content: result.text }
       ];
 
       // Run verification checks (disclaimer, hallucination detection, scope validation)
@@ -318,15 +340,16 @@ export class AiService {
         response: responseText,
         toolCalls,
         verificationChecks: checks,
-        conversationHistory: updatedHistory
+        conversationHistory: updatedHistory,
+        traceId
       };
     } catch (error) {
-      this.logger.error("Agent chat error:", error);
+      this.logger.error('Agent chat error:', error);
 
-      if (error?.message?.includes("API key")) {
+      if (error?.message?.includes('API key')) {
         return {
           response:
-            "The AI service is not properly configured. Please check your API key settings.",
+            'The AI service is not properly configured. Please check your API key settings.',
           toolCalls: [],
           conversationHistory: messages
         };
@@ -334,10 +357,170 @@ export class AiService {
 
       return {
         response:
-          "I encountered an issue processing your request. Please try again later.",
+          'I encountered an issue processing your request. Please try again later.',
         toolCalls: [],
         conversationHistory: messages
       };
     }
   }
+
+  public async agentChatStream({
+    conversationHistory,
+    message,
+    impersonationId,
+    userCurrency,
+    userId,
+    onChunk,
+    onDone,
+    onError
+  }: {
+    conversationHistory?: CoreMessage[];
+    message: string;
+    impersonationId?: string;
+    userCurrency: string;
+    userId: string;
+    onChunk: (text: string) => void;
+    onDone: (metadata: {
+      response: string;
+      toolCalls: any[];
+      verificationChecks: any[];
+      conversationHistory: CoreMessage[];
+      traceId: string;
+    }) => void;
+    onError: (error: string) => void;
+  }) {
+    const messages: CoreMessage[] = [
+      ...(conversationHistory ?? []),
+      { role: 'user' as const, content: message }
+    ];
+
+    const traceId = randomUUID();
+
+    try {
+      const { anthropic, tools } = this.buildAgentConfig({
+        userId,
+        impersonationId,
+        userCurrency
+      });
+
+      const result = streamText({
+        model: anthropic('claude-haiku-4-5-20251001'),
+        system: getAgentSystemPrompt(),
+        tools,
+        toolChoice: 'auto',
+        messages,
+        maxSteps: 10,
+        experimental_telemetry: {
+          isEnabled: true,
+          functionId: 'ghostfolio-ai-agent-stream',
+          metadata: { userId, traceId }
+        }
+      });
+
+      let fullText = '';
+
+      for await (const chunk of result.textStream) {
+        fullText += chunk;
+        onChunk(chunk);
+      }
+
+      const stepsResult = await result.steps;
+
+      const toolCalls = stepsResult
+        .flatMap((step) => step.toolCalls ?? [])
+        .map((tc) => ({
+          toolName: tc.toolName,
+          args: tc.args
+        }));
+
+      const toolResults = stepsResult.flatMap((step) => step.toolResults ?? []);
+
+      const { responseText, checks } = runVerificationChecks({
+        responseText: fullText,
+        toolResults,
+        toolCalls
+      });
+
+      // If verification added extra text (e.g. disclaimer), send the difference
+      if (responseText.length > fullText.length) {
+        onChunk(responseText.slice(fullText.length));
+      }
+
+      const updatedHistory: CoreMessage[] = [
+        ...messages,
+        { role: 'assistant' as const, content: responseText }
+      ];
+
+      onDone({
+        response: responseText,
+        toolCalls,
+        verificationChecks: checks,
+        conversationHistory: updatedHistory,
+        traceId
+      });
+    } catch (error) {
+      this.logger.error('Agent stream error:', error);
+      onError(
+        error?.message?.includes('API key')
+          ? 'The AI service is not properly configured.'
+          : 'I encountered an issue processing your request.'
+      );
+    }
+  }
+
+  public async submitFeedback({
+    traceId,
+    value,
+    userId
+  }: {
+    traceId: string;
+    value: number;
+    userId: string;
+  }) {
+    const langfuseSecretKey = process.env.LANGFUSE_SECRET_KEY;
+    const langfusePublicKey = process.env.LANGFUSE_PUBLIC_KEY;
+    const langfuseBaseUrl =
+      process.env.LANGFUSE_BASEURL || 'https://cloud.langfuse.com';
+
+    if (!langfuseSecretKey || !langfusePublicKey) {
+      this.logger.warn('Langfuse keys not configured — feedback not recorded');
+      return { success: false, reason: 'Langfuse not configured' };
+    }
+
+    try {
+      const credentials = Buffer.from(
+        `${langfusePublicKey}:${langfuseSecretKey}`
+      ).toString('base64');
+
+      const res = await fetch(`${langfuseBaseUrl}/api/public/scores`, {
+        method: 'POST',
+        headers: {
+          'Content-Type': 'application/json',
+          Authorization: `Basic ${credentials}`
+        },
+        body: JSON.stringify({
+          traceId,
+          name: 'user-feedback',
+          value,
+          comment: value === 1 ? 'thumbs up' : 'thumbs down',
+          source: 'API',
+          metadata: { userId }
+        })
+      });
+
+      if (!res.ok) {
+        const errorBody = await res.text();
+        this.logger.warn(
+          `Langfuse score API error: ${res.status} ${errorBody}`
+        );
+        return { success: false, reason: `Langfuse API error: ${res.status}` };
+      }
+
+      this.logger.log(`Feedback recorded: traceId=${traceId} value=${value}`);
+      return { success: true };
+    } catch (error) {
+      this.logger.error('Failed to submit feedback to Langfuse:', error);
+      return { success: false, reason: 'Failed to contact Langfuse' };
+    }
+  }
 }

apps/api/src/app/endpoints/ai/eval/eval.ts

@@ -22,6 +22,7 @@
  *   CATEGORY=<name>     — run only one category (happy_path, edge_case, adversarial, multi_step)
  */
 
+import "dotenv/config";
 import * as http from "http";
 import * as fs from "fs";
 

apps/api/src/services/cron/cron.service.ts

@@ -35,6 +35,23 @@ export class CronService implements OnApplicationBootstrap {
       this.logger.log('Triggering initial data gathering on startup...');
       await this.dataGatheringService.gather7Days();
     }
+
+    // Reload exchange rates after data gathering queue has had time
+    // to process high-priority currency pair jobs (~1 job per 4s).
+    // This ensures rates are populated on fresh Railway deployments.
+    setTimeout(
+      async () => {
+        try {
+          await this.exchangeRateDataService.loadCurrencies();
+          this.logger.log(
+            'Exchange rates reloaded after startup data gathering'
+          );
+        } catch (error) {
+          this.logger.warn('Failed to reload exchange rates on startup', error);
+        }
+      },
+      5 * 60 * 1000
+    );
   }
 
   @Cron(CronService.EVERY_HOUR_AT_RANDOM_MINUTE)

apps/client/src/app/pages/ai-chat/ai-chat-page.component.ts

@@ -1,3 +1,4 @@
+import { TokenStorageService } from '@ghostfolio/client/services/token-storage.service';
 import { DataService } from '@ghostfolio/ui/services';
 
 import { CommonModule } from '@angular/common';
@@ -6,11 +7,13 @@ import {
   ChangeDetectorRef,
   Component,
   ElementRef,
+  NgZone,
   OnDestroy,
   ViewChild
 } from '@angular/core';
 import { FormsModule } from '@angular/forms';
 import { MatButtonModule } from '@angular/material/button';
+import { MatIconModule } from '@angular/material/icon';
 import { MatInputModule } from '@angular/material/input';
 import { MatProgressSpinnerModule } from '@angular/material/progress-spinner';
 import { Subject, takeUntil } from 'rxjs';
@@ -18,6 +21,8 @@ import { Subject, takeUntil } from 'rxjs';
 interface ChatMessage {
   role: 'user' | 'assistant';
   content: string;
+  traceId?: string;
+  feedback?: 1 | -1;
 }
 
 @Component({
@@ -27,6 +32,7 @@ interface ChatMessage {
     CommonModule,
     FormsModule,
     MatButtonModule,
+    MatIconModule,
     MatInputModule,
     MatProgressSpinnerModule
   ],
@@ -43,11 +49,14 @@ export class GfAiChatPageComponent implements OnDestroy {
   public isLoading = false;
 
   private conversationHistory: any[] = [];
+  private abortController: AbortController | null = null;
   private unsubscribeSubject = new Subject<void>();
 
   public constructor(
     private changeDetectorRef: ChangeDetectorRef,
-    private dataService: DataService
+    private dataService: DataService,
+    private ngZone: NgZone,
+    private tokenStorageService: TokenStorageService
   ) {}
 
   public sendMessage() {
@@ -62,46 +71,134 @@ export class GfAiChatPageComponent implements OnDestroy {
     this.changeDetectorRef.markForCheck();
     this.scrollToBottom();
 
+    const assistantMsg: ChatMessage = { role: 'assistant', content: '' };
+    this.messages.push(assistantMsg);
+    this.changeDetectorRef.markForCheck();
+
+    this.streamResponse(message, assistantMsg);
+  }
+
+  public submitFeedback(msg: ChatMessage, value: 1 | -1) {
+    if (!msg.traceId || msg.feedback) {
+      return;
+    }
+
+    msg.feedback = value;
+    this.changeDetectorRef.markForCheck();
+
     this.dataService
-      .postAgentChat({
+      .postAgentFeedback({ traceId: msg.traceId, value })
+      .pipe(takeUntil(this.unsubscribeSubject))
+      .subscribe();
+  }
+
+  public onKeyDown(event: KeyboardEvent) {
+    if (event.key === 'Enter' && !event.shiftKey) {
+      event.preventDefault();
+      this.sendMessage();
+    }
+  }
+
+  public ngOnDestroy() {
+    this.abortController?.abort();
+    this.unsubscribeSubject.next();
+    this.unsubscribeSubject.complete();
+  }
+
+  private async streamResponse(message: string, assistantMsg: ChatMessage) {
+    this.abortController = new AbortController();
+    const token = this.tokenStorageService.getToken();
+
+    try {
+      const res = await fetch('/api/v1/ai/agent/stream', {
+        method: 'POST',
+        headers: {
+          'Content-Type': 'application/json',
+          ...(token ? { Authorization: `Bearer ${token}` } : {})
+        },
+        body: JSON.stringify({
           message,
           conversationHistory: this.conversationHistory
-      })
+        }),
-      .pipe(takeUntil(this.unsubscribeSubject))
+        signal: this.abortController.signal
-      .subscribe({
-        next: (response) => {
-          this.messages.push({
-            role: 'assistant',
-            content: response.response
       });
-          this.conversationHistory = response.conversationHistory;
+
-          this.isLoading = false;
+      if (!res.ok || !res.body) {
+        throw new Error(`HTTP ${res.status}`);
+      }
+
+      const reader = res.body.getReader();
+      const decoder = new TextDecoder();
+      let buffer = '';
+
+      while (true) {
+        const { done, value } = await reader.read();
+
+        if (done) {
+          break;
+        }
+
+        buffer += decoder.decode(value, { stream: true });
+
+        // Parse SSE events from buffer
+        const lines = buffer.split('\n');
+        buffer = lines.pop() || '';
+
+        let eventType = '';
+
+        for (const line of lines) {
+          if (line.startsWith('event: ')) {
+            eventType = line.slice(7).trim();
+          } else if (line.startsWith('data: ')) {
+            const data = line.slice(6);
+
+            this.ngZone.run(() => {
+              if (eventType === 'text') {
+                assistantMsg.content += JSON.parse(data);
                 this.changeDetectorRef.markForCheck();
                 this.scrollToBottom();
-        },
+              } else if (eventType === 'done') {
-        error: () => {
+                const metadata = JSON.parse(data);
-          this.messages.push({
+                assistantMsg.traceId = metadata.traceId;
-            role: 'assistant',
+                assistantMsg.content = metadata.response;
-            content:
+                this.conversationHistory = metadata.conversationHistory;
-              'Sorry, something went wrong. Please try again.'
-          });
                 this.isLoading = false;
                 this.changeDetectorRef.markForCheck();
                 this.scrollToBottom();
+              } else if (eventType === 'error') {
+                const errorData = JSON.parse(data);
+                assistantMsg.content =
+                  errorData.error ||
+                  'Sorry, something went wrong. Please try again.';
+                this.isLoading = false;
+                this.changeDetectorRef.markForCheck();
               }
             });
+
+            eventType = '';
+          }
+        }
       }
 
-  public onKeyDown(event: KeyboardEvent) {
+      // If stream ended without a done event, finalize
-    if (event.key === 'Enter' && !event.shiftKey) {
+      if (this.isLoading) {
-      event.preventDefault();
+        this.ngZone.run(() => {
-      this.sendMessage();
+          this.isLoading = false;
+          this.changeDetectorRef.markForCheck();
+        });
       }
+    } catch (error: any) {
+      if (error?.name === 'AbortError') {
+        return;
       }
 
-  public ngOnDestroy() {
+      this.ngZone.run(() => {
-    this.unsubscribeSubject.next();
+        assistantMsg.content = 'Sorry, something went wrong. Please try again.';
-    this.unsubscribeSubject.complete();
+        this.isLoading = false;
+        this.changeDetectorRef.markForCheck();
+        this.scrollToBottom();
+      });
+    }
   }
 
   private scrollToBottom() {

apps/client/src/app/pages/ai-chat/ai-chat-page.html

@@ -5,12 +5,19 @@
     <div #messagesContainer class="messages-container">
       @if (messages.length === 0 && !isLoading) {
         <div class="empty-state">
-          <p i18n>Ask me anything about your portfolio, holdings, performance, or market data.</p>
+          <p i18n>
+            Ask me anything about your portfolio, holdings, performance, or
+            market data.
+          </p>
         </div>
       }
 
       @for (msg of messages; track $index) {
-        <div class="message" [class.user]="msg.role === 'user'" [class.assistant]="msg.role === 'assistant'">
+        <div
+          class="message"
+          [class.assistant]="msg.role === 'assistant'"
+          [class.user]="msg.role === 'user'"
+        >
           <div class="message-label">
             @if (msg.role === 'user') {
               <strong i18n>You</strong>
@@ -18,36 +25,59 @@
               <strong i18n>Assistant</strong>
             }
           </div>
+          @if (msg.role === 'assistant' && !msg.content && isLoading) {
+            <mat-spinner diameter="24"></mat-spinner>
+          } @else {
             <div class="message-content">{{ msg.content }}</div>
+          }
+          @if (msg.role === 'assistant' && msg.traceId) {
+            <div class="feedback-buttons">
+              <button
+                class="feedback-btn"
+                mat-icon-button
+                title="Helpful"
+                [class.selected]="msg.feedback === 1"
+                [disabled]="!!msg.feedback"
+                (click)="submitFeedback(msg, 1)"
+              >
+                <mat-icon>thumb_up</mat-icon>
+              </button>
+              <button
+                class="feedback-btn"
+                mat-icon-button
+                title="Not helpful"
+                [class.selected]="msg.feedback === -1"
+                [disabled]="!!msg.feedback"
+                (click)="submitFeedback(msg, -1)"
+              >
+                <mat-icon>thumb_down</mat-icon>
+              </button>
             </div>
           }
-
-      @if (isLoading) {
-        <div class="message assistant">
-          <div class="message-label"><strong i18n>Assistant</strong></div>
-          <mat-spinner diameter="24"></mat-spinner>
         </div>
       }
     </div>
 
     <div class="input-area">
-      <mat-form-field class="w-100" appearance="outline">
+      <mat-form-field appearance="outline" class="w-100">
         <input
+          i18n-placeholder
           matInput
           placeholder="Ask about your portfolio..."
-          i18n-placeholder
+          [disabled]="isLoading"
           [(ngModel)]="userInput"
           (keydown)="onKeyDown($event)"
-          [disabled]="isLoading"
         />
       </mat-form-field>
       <button
-        mat-flat-button
         color="primary"
+        i18n
+        mat-flat-button
         [disabled]="!userInput.trim() || isLoading"
         (click)="sendMessage()"
-        i18n
+      >
-      >Send</button>
+        Send
+      </button>
     </div>
   </div>
 </div>

apps/client/src/app/pages/ai-chat/ai-chat-page.scss

@@ -71,6 +71,38 @@
   line-height: 1.5;
 }
 
+.feedback-buttons {
+  display: flex;
+  gap: 0.25rem;
+  margin-top: 0.25rem;
+
+  .feedback-btn {
+    width: 28px;
+    height: 28px;
+    line-height: 28px;
+    opacity: 0.5;
+
+    mat-icon {
+      font-size: 16px;
+      width: 16px;
+      height: 16px;
+    }
+
+    &:hover:not(:disabled) {
+      opacity: 1;
+    }
+
+    &.selected {
+      opacity: 1;
+      color: rgb(var(--palette-primary-500));
+    }
+
+    &:disabled:not(.selected) {
+      opacity: 0.3;
+    }
+  }
+}
+
 .input-area {
   display: flex;
   gap: 0.5rem;

libs/ui/src/lib/services/data.service.ts

@@ -740,17 +740,19 @@ export class DataService {
     return this.http.get<WatchlistResponse>('/api/v1/watchlist');
   }
 
-  public postAgentChat(body: {
+  public postAgentChat(body: { message: string; conversationHistory?: any[] }) {
-    message: string;
-    conversationHistory?: any[];
-  }) {
     return this.http.post<{
       response: string;
       toolCalls: Array<{ toolName: string; args: any }>;
       conversationHistory: any[];
+      traceId?: string;
     }>('/api/v1/ai/agent', body);
   }
 
+  public postAgentFeedback(body: { traceId: string; value: number }) {
+    return this.http.post<{ success: boolean }>('/api/v1/ai/feedback', body);
+  }
+
   public loginAnonymous(accessToken: string) {
     return this.http.post<OAuthResponse>('/api/v1/auth/anonymous', {
       accessToken