Merge branch 'main' of https://github.com/kaiban-ai/KaibanJS into workflow-driven-agent

Author: anthonydevs17

playground/react/package-lock.json

@@ -28,6 +28,7 @@
     "langchain": "^0.3.33",
     "mathjs": "^13.0.3",
     "octokit": "^4.0.2",
+    "openai": "^6.9.0",
     "react": "^18.2.0",
     "react-dom": "^18.2.0",
     "react-markdown": "^10.1.0"

playground/react/package.json

@@ -0,0 +1,17 @@
+import AgentsBoardDebugger from '../AgentsBoardDebugger';
+import teamOpenAI from '../teams/conference_analysis/openai';
+import '../index.css';
+
+// More on how to set up stories at: https://storybook.js.org/docs/writing-stories#default-export
+export default {
+  title: 'Teams/Conference Analysis Team',
+  component: AgentsBoardDebugger,
+};
+
+// More on writing stories with args: https://storybook.js.org/docs/writing-stories/args
+export const withOpenAI = {
+  args: {
+    team: teamOpenAI,
+    title: 'Conference Analysis with OpenAI Whisper',
+  },
+};

playground/react/src/stories/ConferenceAnalysisTeam.stories.js

@@ -0,0 +1,255 @@
+import { Agent, Task, Team } from 'kaibanjs';
+import { z } from 'zod';
+import OpenAI from 'openai';
+
+// ──── Custom Tool: Audio Transcription ────────────────────────────────
+/**
+ * Custom tool for transcribing audio files using OpenAI Whisper API
+ * Downloads audio from URL and transcribes it client-side
+ */
+class AudioTranscriptionTool {
+  name = 'audio_transcription';
+
+  description =
+    'Downloads an audio file from a URL and transcribes it using OpenAI Whisper API. Returns the full transcription text.';
+
+  schema = z.object({
+    audioUrl: z.string().describe('URL of the audio file to transcribe'),
+  });
+
+  openaiClient = null;
+
+  constructor(apiKey) {
+    if (!apiKey) {
+      throw new Error('API key is required');
+    }
+    this.openaiClient = new OpenAI({
+      apiKey,
+      dangerouslyAllowBrowser: true, // Required for browser usage
+    });
+  }
+
+  async invoke(input) {
+    const { audioUrl } = input;
+
+    try {
+      // Download audio file from URL
+      const response = await fetch(audioUrl);
+      if (!response.ok) {
+        throw new Error(`Failed to download audio: ${response.statusText}`);
+      }
+
+      const audioBlob = await response.blob();
+
+      // Convert blob to File object for OpenAI API
+      const audioFile = new File([audioBlob], 'audio.mp3', {
+        type: audioBlob.type || 'audio/mpeg',
+      });
+
+      // Transcribe using OpenAI Whisper
+      const transcription = await this.openaiClient.audio.transcriptions.create(
+        {
+          file: audioFile,
+          model: 'gpt-4o-mini-transcribe',
+          response_format: 'text',
+          //   model: 'gpt-4o-transcribe-diarize',
+          //   response_format: 'diarized_json',
+          //   chunking_strategy: 'auto',
+        }
+      );
+      console.log('transcription', transcription);
+
+      return typeof transcription === 'string'
+        ? transcription
+        : JSON.stringify(transcription);
+    } catch (error) {
+      return JSON.stringify({
+        status: 'ERROR',
+        message: `Failed to transcribe audio: ${
+          error instanceof Error ? error.message : 'Unknown error'
+        }`,
+      });
+    }
+  }
+}
+
+// Initialize transcription tool with API key from environment
+const transcriptionTool = new AudioTranscriptionTool(
+  import.meta.env.VITE_OPENAI_API_KEY
+);
+
+// ──── Agents ───────────────────────────────────────────────────────────
+
+/**
+ * Transcription Agent
+ * Handles audio transcription using the custom tool
+ */
+const transcriber = new Agent({
+  name: 'Transcriber',
+  role: 'Audio Transcription Specialist',
+  goal: 'Transcribe audio recordings to accurate text ',
+  background: 'Expert in audio processing and speech-to-text conversion.',
+  type: 'ReactChampionAgent',
+  tools: [transcriptionTool],
+});
+
+/**
+ * Content Analyst Agent
+ * Analyzes transcription to extract topics, context, participants, and generate summary
+ * This agent handles multiple related analysis tasks
+ */
+const contentAnalyst = new Agent({
+  name: 'Analyst',
+  role: 'Content Analysis Specialist',
+  goal: 'Analyze conference content to identify topics, context, participants, and create summaries.',
+  background:
+    'Expert in content analysis, information extraction, and summarization with experience in meeting analysis.',
+  type: 'ReactChampionAgent',
+  tools: [],
+});
+
+/**
+ * Information Extractor Agent
+ * Extracts action items and relevant notes from the transcription
+ */
+const informationExtractor = new Agent({
+  name: 'Extractor',
+  role: 'Information Extraction Specialist',
+  goal: 'Extract actionable items, responsibilities, and key notes from conference discussions.',
+  background:
+    'Specialist in identifying action items, tracking responsibilities, and extracting key insights from conversations.',
+  type: 'ReactChampionAgent',
+  tools: [],
+});
+
+/**
+ * Document Consolidator Agent
+ * Consolidates all analysis into a final markdown document
+ */
+const consolidator = new Agent({
+  name: 'Consolidator',
+  role: 'Document Consolidation Specialist',
+  goal: 'Consolidate all conference analysis into a comprehensive, well-structured markdown document.',
+  background:
+    'Expert in document creation and information synthesis, creating professional meeting notes and reports.',
+  type: 'ReactChampionAgent',
+  tools: [],
+});
+
+// ──── Tasks ─────────────────────────────────────────────────────────────
+
+/**
+ * Task 1: Transcribe audio
+ */
+const transcriptionTask = new Task({
+  description: `Transcribe the audio file from the provided URL: {audioUrl}`,
+  expectedOutput:
+    'Complete transcription of the conference audio in plain text format.',
+  agent: transcriber,
+});
+
+/**
+ * Task 2: Analyze topics and context
+ * Uses contentAnalyst agent
+ */
+const topicContextTask = new Task({
+  description: `Analyze the transcription to identify the main topics discussed and the overall context of the conference. 
+    Provide a clear overview of what the conference was about and the key themes.`,
+  expectedOutput:
+    'Detailed analysis of main topics and context of the conference.',
+  agent: contentAnalyst,
+});
+
+/**
+ * Task 3: Extract participants
+ * Uses contentAnalyst agent
+ */
+const participantsTask = new Task({
+  description: `Identify all participants mentioned in the transcription. 
+    Extract their names, roles, titles, and any other relevant information about them.`,
+  expectedOutput:
+    'Complete list of participants with their details (names, roles, titles, etc.).',
+  agent: contentAnalyst,
+});
+
+/**
+ * Task 4: Generate summary
+ * Uses contentAnalyst agent
+ */
+const summaryTask = new Task({
+  description: `Create a concise and comprehensive summary of the conference based on the transcription. 
+    Highlight the main points, decisions made, and important discussions.`,
+  expectedOutput:
+    'A well-structured summary of the conference covering main points and decisions.',
+  agent: contentAnalyst,
+});
+
+/**
+ * Task 5: Extract action items
+ * Uses informationExtractor agent
+ */
+const actionItemsTask = new Task({
+  description: `Identify all action items mentioned in the conference. 
+    For each action item, extract the task description and the person responsible (if mentioned).`,
+  expectedOutput:
+    'List of action items with descriptions and assigned responsible parties.',
+  agent: informationExtractor,
+});
+
+/**
+ * Task 6: Extract relevant notes
+ * Uses informationExtractor agent
+ */
+const notesTask = new Task({
+  description: `Extract and organize all relevant notes, insights, and important information from the conference. 
+    Focus on key takeaways, decisions, and valuable insights that should be documented.`,
+  expectedOutput:
+    'Organized collection of relevant notes, insights, and key takeaways from the conference.',
+  agent: informationExtractor,
+});
+
+/**
+ * Task 7: Consolidate final document
+ * Uses consolidator agent - This is the deliverable
+ */
+const consolidationTask = new Task({
+  description: `Consolidate all the analysis results into a comprehensive markdown document. 
+    Include:
+    - Conference Summary
+    - Topics & Context
+    - Participants (with details)
+    - Action Items (with responsible parties)
+    - Key Notes & Insights
+    
+    Format everything in a clear, professional markdown structure suitable for meeting notes.`,
+  expectedOutput:
+    'A complete, well-formatted markdown document containing all conference analysis results.',
+  isDeliverable: true,
+  agent: consolidator,
+});
+
+// ──── Team ──────────────────────────────────────────────────────────────
+
+// Create the team
+const team = new Team({
+  name: 'Conference Analysis Team',
+  agents: [transcriber, contentAnalyst, informationExtractor, consolidator],
+  tasks: [
+    transcriptionTask,
+    topicContextTask,
+    participantsTask,
+    summaryTask,
+    actionItemsTask,
+    notesTask,
+    consolidationTask,
+  ],
+  inputs: {
+    audioUrl:
+      'https://raw.githubusercontent.com/ringcentral/ringcentral-api-docs/main/resources/RC_Conversation_Sample.mp3',
+  },
+  env: {
+    OPENAI_API_KEY: import.meta.env.VITE_OPENAI_API_KEY,
+  },
+});
+
+export default team;