apexbot 1.0.1 → 1.0.4

package/README.md

@@ -1,6 +1,6 @@
-# 🦊 ApexBot — Your Free, Private AI Assistant
+# ApexBot
 
-<div align="center">
+Personal AI assistant you run on your own devices. Free with Ollama (local AI).
 
 ```
     ___    ____  _______  ______  ____  ______
@@ -9,112 +9,119 @@
  / ___ |/ ____/ /___ /   / /_/ / /_/ / / /    
 /_/  |_/_/   /_____//_/|_/_____/\____/ /_/    
                                                
-       🦊 Your Free AI Assistant 🦊
+       Your Free AI Assistant
 ```
 
 [![CI](https://github.com/YOUR_USERNAME/apexbot/actions/workflows/ci.yml/badge.svg)](https://github.com/YOUR_USERNAME/apexbot/actions)
-[![npm version](https://badge.fury.io/js/apexbot.svg)](https://www.npmjs.com/package/apexbot)
-[![MIT License](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)
+[![npm package](https://badge.fury.io/js/apexbot.svg)](https://www.npmjs.com/package/apexbot)
+[![license](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)
 [![TypeScript](https://img.shields.io/badge/TypeScript-5.0-blue.svg)](https://www.typescriptlang.org/)
 
-**100% Free • Open Source • Private • Self-Hosted**
+100% Free | Open Source | Private | Self-Hosted
 
-[Installation](#-installation) • [Features](#-features) • [Quick Start](#-quick-start) • [Commands](#-commands) • [Contributing](#-contributing)
-
-</div>
+[Installation](#installation) |
+[Features](#features) |
+[Quick Start](#quick-start) |
+[Commands](#commands) |
+[Contributing](#contributing)
 
 ---
 
-## ⚡ Installation
+## Installation
+
+Runtime: Node 18+
+
+### npm (recommended)
+
+```bash
+npm install -g apexbot
+apexbot onboard
+```
+
+### One-line install
 
-### One-Line Install
+Windows (PowerShell):
 
-**Windows (PowerShell):**
 ```powershell
 iwr -useb https://raw.githubusercontent.com/YOUR_USERNAME/apexbot/main/scripts/install.ps1 | iex
 ```
 
-**macOS/Linux:**
-```bash
-curl -fsSL https://raw.githubusercontent.com/YOUR_USERNAME/apexbot/main/scripts/install.sh | bash
-```
+macOS/Linux:
 
-**npm (global):**
 ```bash
-npm install -g apexbot
-apexbot onboard
+curl -fsSL https://raw.githubusercontent.com/YOUR_USERNAME/apexbot/main/scripts/install.sh | bash
 ```
 
-### Manual Install
+### From source
 
 ```bash
-# Clone repository
 git clone https://github.com/YOUR_USERNAME/apexbot.git
 cd apexbot
 npm install
 npm run build
-
-# Run setup wizard
 npm run onboard
 ```
 
 ---
 
-## ✨ Features
-
-### 🤖 100% Free AI with Ollama
-- **No API keys needed** — runs completely locally
-- **No cloud costs** — your hardware, your AI
-- **Private** — conversations never leave your machine
-- **Multiple models** — llama3.2, mistral, codellama, qwen2.5
-
-### 📱 Multi-Channel Support
-| Channel | Status | Description |
-|---------|--------|-------------|
-| **Telegram** | ✅ Ready | Full bot integration |
-| **Discord** | ✅ Ready | Guilds and DMs |
-| **WebChat** | ✅ Ready | Built-in web dashboard |
-| **WhatsApp** | 🚧 Soon | Via Baileys |
-
-### 🧠 AI Providers
-| Provider | Cost | Description |
-|----------|------|-------------|
-| **Ollama** | **FREE** | Local AI (recommended!) |
-| Google Gemini | Free tier | Cloud AI |
-| Anthropic Claude | Paid | Cloud AI |
-| OpenAI GPT | Paid | Cloud AI |
-
-### 🛡️ Self-Hosted & Private
-- Your data stays with you
-- No telemetry or tracking
-- Full control over your bot
+## Features
+
+### Local AI with Ollama
+
+ApexBot runs AI models locally on your machine using Ollama. No API keys, no cloud costs, complete privacy.
+
+- **Free**: No subscriptions or usage fees
+- **Private**: Conversations never leave your computer
+- **Offline**: Works without internet after model download
+- **Models**: llama3.2, mistral, codellama, qwen2.5, and more
+
+### Multi-channel support
+
+| Channel    | Status   | Description            |
+|------------|----------|------------------------|
+| Telegram   | Ready    | Full bot integration   |
+| Discord    | Ready    | Guilds and DMs         |
+| WebChat    | Ready    | Built-in web dashboard |
+| WhatsApp   | Planned  | Via Baileys            |
+
+### AI providers
+
+| Provider         | Cost    | Notes                      |
+|------------------|---------|----------------------------|
+| **Ollama**       | Free    | Local AI (recommended)     |
+| Google Gemini    | Free tier | Cloud API               |
+| Anthropic Claude | Paid    | Cloud API                  |
+| OpenAI GPT       | Paid    | Cloud API                  |
 
 ---
 
-## 🚀 Quick Start
+## Quick Start
 
-### 1. Install Ollama (for free AI)
+### 1. Install Ollama
+
+Windows:
 
-**Windows:**
 ```powershell
 winget install Ollama.Ollama
 ```
 
-**macOS:**
+macOS:
+
 ```bash
 brew install ollama
 ```
 
-**Linux:**
+Linux:
+
 ```bash
 curl -fsSL https://ollama.com/install.sh | sh
 ```
 
-### 2. Pull a model
+### 2. Download a model
 
 ```bash
 ollama pull llama3.2
-ollama serve  # Keep running in background
+ollama serve
 ```
 
 ### 3. Install ApexBot
@@ -123,38 +130,35 @@ ollama serve  # Keep running in background
 npm install -g apexbot
 ```
 
-### 4. Run setup wizard
+### 4. Run setup
 
 ```bash
 apexbot onboard
 ```
 
-The wizard will guide you through:
-- Choosing AI provider (Ollama recommended!)
-- Setting up Telegram/Discord
-- Configuring your bot
+The wizard walks you through AI provider selection, channel setup (Telegram/Discord), and configuration.
 
-### 5. Start your bot
+### 5. Start the bot
 
 ```bash
 apexbot daemon start
 ```
 
-Done! 🎉 Your bot is now running.
+Your assistant is now running in the background.
 
 ---
 
-## 🎮 Commands
+## Commands
 
-### CLI Commands
+### CLI
 
 ```bash
-# Setup & Configuration
+# Setup
 apexbot onboard            # Interactive setup wizard
 apexbot config             # Show configuration
 apexbot config --reset     # Reset configuration
 
-# Daemon (background service)
+# Daemon (background)
 apexbot daemon start       # Start in background
 apexbot daemon stop        # Stop daemon
 apexbot daemon restart     # Restart daemon
@@ -162,85 +166,73 @@ apexbot daemon status      # Check status
 
 # Gateway (foreground)
 apexbot gateway            # Start gateway server
-apexbot gateway --verbose  # With debug logs
+apexbot gateway --verbose  # With debug output
 
-# Other
+# Utilities
 apexbot status             # Show status
 apexbot models             # Manage Ollama models
 ```
 
-### Chat Commands
+### Chat commands
 
-| Command | Description |
-|---------|-------------|
-| `/help` | Show available commands |
-| `/status` | Show session info |
-| `/new` | Start new conversation |
-| `/model <name>` | Switch AI model |
+| Command        | Description              |
+|----------------|--------------------------|
+| `/help`        | Show available commands  |
+| `/status`      | Show session info        |
+| `/new`         | Start new conversation   |
+| `/model <name>`| Switch AI model          |
 
 ---
 
-## 🏗️ Architecture
+## Architecture
 
 ```
-┌─────────────────────────────────────────────────────────────┐
-│                    Messaging Channels                       │
-│         Telegram │ Discord │ WebChat │ WhatsApp            │
-└────────────┬─────────┬─────────┬─────────┬──────────────────┘
-             │         │         │         │
-             ▼         ▼         ▼         ▼
-┌─────────────────────────────────────────────────────────────┐
-│                     Gateway Server                          │
-│              (HTTP + WebSocket on port 18789)               │
-├─────────────────────────────────────────────────────────────┤
-│  Sessions Manager  │  Event Bus  │  Rate Limiter           │
-└────────────────────┼─────────────┼──────────────────────────┘
-                     │             │
-                     ▼             ▼
-┌─────────────────────────────────────────────────────────────┐
-│                     AI Agent Manager                        │
-│         Ollama │ Gemini │ Claude │ OpenAI                  │
-└─────────────────────────────────────────────────────────────┘
+Messaging Channels
+Telegram | Discord | WebChat | WhatsApp
+         |         |         |
+         v         v         v
++----------------------------------------+
+|            Gateway Server              |
+|       (HTTP + WebSocket :18789)        |
++----------------------------------------+
+|  Sessions  |  Event Bus  |  Rate Limit |
++----------------------------------------+
+                   |
+                   v
++----------------------------------------+
+|           AI Agent Manager             |
+|    Ollama | Gemini | Claude | OpenAI   |
++----------------------------------------+
 ```
 
 ---
 
-## 🆚 ApexBot vs Clawdbot
-
-| Feature | Clawdbot | ApexBot |
-|---------|----------|---------|
-| **Cost** | Cloud APIs (paid) | **100% FREE with Ollama** |
-| **Privacy** | Cloud-dependent | **Fully local & private** |
-| **Channels** | WhatsApp, Telegram, Discord, iMessage | Telegram, Discord, WebChat |
-| **AI** | Claude, GPT | **Ollama (free!)**, Gemini, Claude, GPT |
-| **Setup** | Complex | **Simple wizard** |
-| **License** | Proprietary | **MIT (Open Source)** |
-
----
-
-## 📁 Project Structure
+## Project Structure
 
 ```
 apexbot/
-├── src/
-│   ├── cli/              # Command-line interface
-│   ├── gateway/          # HTTP/WebSocket server & dashboard
-│   ├── channels/         # Telegram, Discord, WebChat
-│   ├── agent/            # AI provider integrations
-│   ├── sessions/         # Conversation management
-│   └── core/             # Event bus, utilities
-├── scripts/
-│   ├── install.ps1       # Windows installer
-│   └── install.sh        # macOS/Linux installer
-├── package.json
-└── README.md
+  src/
+    adapters/       Provider integrations (Manifold, etc.)
+    agent/          AI agent manager and message handling
+    backtest/       Backtesting framework
+    channels/       Channel adapters (Telegram, Discord)
+    cli/            Command-line interface
+    core/           Event bus and shared utilities
+    gateway/        HTTP/WebSocket server and dashboard
+    math/           Expected value and Kelly criterion
+    safety/         Rate limiting and content filtering
+    sessions/       Conversation state management
+    strategy/       Trading strategies (arbitrage)
+  scripts/
+    install.ps1     Windows installer
+    install.sh      Unix installer
 ```
 
 ---
 
-## 🔧 Configuration
+## Configuration
 
-Configuration is stored in `~/.apexbot/config.json`:
+Configuration lives in `~/.apexbot/config.json`:
 
 ```json
 {
@@ -260,7 +252,7 @@ Configuration is stored in `~/.apexbot/config.json`:
 }
 ```
 
-Or use environment variables:
+Environment variables also work:
 
 ```bash
 export OLLAMA_URL=http://localhost:11434
@@ -270,38 +262,27 @@ export TELEGRAM_BOT_TOKEN=your-token
 
 ---
 
-## 🤝 Contributing
+## Contributing
 
-Contributions are welcome! See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines.
+See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines.
 
 ```bash
-# Development
 git clone https://github.com/YOUR_USERNAME/apexbot.git
 cd apexbot
 npm install
-npm run dev  # Start with hot-reload
+npm run dev
 ```
 
 ---
 
-## 📄 License
+## License
 
-MIT License — see [LICENSE](LICENSE) for details.
+MIT License. See [LICENSE](LICENSE).
 
 ---
 
-## 🙏 Acknowledgments
-
-- Inspired by [Clawdbot](https://github.com/clawdbot/clawdbot)
-- Powered by [Ollama](https://ollama.com) for free local AI
-- Built with TypeScript, grammY, discord.js
-
----
-
-<div align="center">
-
-**Made with ❤️ by the ApexBot community**
-
-⭐ Star this repo if you find it useful!
+## Acknowledgments
 
-</div>
+- Inspired by [Clawdbot](https://github.com/moltbot/moltbot)
+- Powered by [Ollama](https://ollama.com) for local AI
+- Built with TypeScript, grammy, discord.js

package/dist/agent/agentManager.js

@@ -7,6 +7,8 @@
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.AgentManager = void 0;
 const generative_ai_1 = require("@google/generative-ai");
+const tools_1 = require("../tools");
+const toolExecutor_1 = require("./toolExecutor");
 class AgentManager {
     config = null;
     googleClient = null;
@@ -42,6 +44,7 @@ You are running locally on the user's machine. No data leaves their computer. Yo
             maxTokens: 4096,
             temperature: 0.7,
             systemPrompt: this.defaultSystemPrompt,
+            enableTools: true, // Tools enabled by default
             ...config,
         };
         // Initialize client based on provider
@@ -65,9 +68,10 @@ You are running locally on the user's machine. No data leaves their computer. Yo
     async process(session, message) {
         if (!this.config) {
             console.warn('[Agent] Not configured');
-            return { text: '⚠️ Agent not configured. Please set up an AI provider.' };
+            return { text: 'Agent not configured. Please set up an AI provider.' };
         }
         const userText = message.text || '';
+        console.log(`[Agent] Processing message: ${userText.slice(0, 50)}...`);
         // Handle slash commands
         if (userText.startsWith('/')) {
             return this.handleCommand(session, userText);
@@ -93,8 +97,45 @@ You are running locally on the user's machine. No data leaves their computer. Yo
                     response = await this.processWithKimi(history);
                     break;
                 default:
-                    response = { text: '❌ Unknown AI provider' };
+                    response = { text: 'Unknown AI provider' };
             }
+            // Check for tool calls in response
+            if (this.config.enableTools) {
+                const toolCalls = (0, toolExecutor_1.parseToolCalls)(response.text);
+                if (toolCalls.length > 0) {
+                    console.log(`[Agent] Found ${toolCalls.length} tool calls`);
+                    // Build tool context
+                    const context = (0, toolExecutor_1.buildToolContext)(session.id, message.userId || 'unknown', message.channel, { workspaceDir: process.cwd() });
+                    // Execute tools
+                    const toolResults = await (0, toolExecutor_1.executeToolCalls)(toolCalls, context);
+                    response.toolCalls = toolResults;
+                    // Format results and continue conversation
+                    const resultsText = (0, toolExecutor_1.formatToolResults)(toolResults);
+                    if (resultsText) {
+                        // Add tool results to history and get follow-up response
+                        const followUpHistory = [
+                            ...history,
+                            { role: 'assistant', content: response.text, timestamp: Date.now() },
+                            { role: 'user', content: `Tool results:\n${resultsText}\n\nPlease continue based on these results.`, timestamp: Date.now() },
+                        ];
+                        // Get follow-up response
+                        let followUp;
+                        switch (this.config.provider) {
+                            case 'ollama':
+                                followUp = await this.processWithOllama(followUpHistory);
+                                break;
+                            case 'google':
+                                followUp = await this.processWithGemini(followUpHistory);
+                                break;
+                            default:
+                                followUp = { text: resultsText };
+                        }
+                        // Combine responses
+                        response.text = followUp.text;
+                    }
+                }
+            }
+            console.log(`[Agent] Generated response: ${response.text.slice(0, 50)}...`);
             // Save to session
             session.messages.push({ role: 'user', content: userText, timestamp: Date.now() });
             session.messages.push({ role: 'assistant', content: response.text, timestamp: Date.now() });
@@ -103,9 +144,8 @@ You are running locally on the user's machine. No data leaves their computer. Yo
             return response;
         }
         catch (e) {
-            console.error('[Agent] Error:', e);
-            // Escape special characters for Telegram
-            const errorMsg = (e.message || 'Unknown error').slice(0, 200).replace(/[_*\[\]()~`>#+=|{}.!-]/g, '\\$&');
+            console.error('[Agent] Error processing:', e);
+            const errorMsg = (e.message || 'Unknown error').slice(0, 200);
             return { text: `Error: ${errorMsg}` };
         }
     }
@@ -120,7 +160,7 @@ You are running locally on the user's machine. No data leaves their computer. Yo
         const apiKey = cfg.apiKey || process.env.KIMI_API_KEY;
         const apiUrl = cfg.apiUrl || process.env.KIMI_API_URL;
         if (!apiKey || !apiUrl) {
-            return { text: '⚠️ Kimi provider not configured. Set config.apiUrl and apiKey (or KIMI_API_URL/KIMI_API_KEY).' };
+            return { text: 'Kimi provider not configured. Set config.apiUrl and apiKey (or KIMI_API_URL/KIMI_API_KEY).' };
         }
         // Build prompt from history (system prompt + conversation)
         const systemPrompt = history.find(m => m.role === 'system')?.content || '';
@@ -150,7 +190,7 @@ You are running locally on the user's machine. No data leaves their computer. Yo
         const data = await res.json().catch(() => ({}));
         const text = data.text || data.output || (data.choices && data.choices[0]?.text) || '';
         return {
-            text: String(text || '').trim() || '⚠️ Kimi returned an empty response.',
+            text: String(text || '').trim() || 'Kimi returned an empty response.',
         };
     }
     async processWithOllama(history) {
@@ -160,6 +200,7 @@ You are running locally on the user's machine. No data leaves their computer. Yo
         const apiUrl = cfg.apiUrl || 'http://localhost:11434';
         const model = cfg.model || 'llama3.2';
         const temperature = cfg.temperature ?? 0.7;
+        console.log(`[Agent] Calling Ollama at ${apiUrl} with model ${model}...`);
         // Build messages array for Ollama chat API
         const messages = history.map(m => ({
             role: m.role,
@@ -183,12 +224,14 @@ You are running locally on the user's machine. No data leaves their computer. Yo
             });
             if (!res.ok) {
                 const err = await res.text().catch(() => res.statusText);
+                console.error(`[Agent] Ollama error: ${err}`);
                 throw new Error(`Ollama error: ${err}`);
             }
             const data = await res.json();
             const text = data.message?.content || '';
+            console.log(`[Agent] Ollama response received: ${text.slice(0, 50)}...`);
             return {
-                text: String(text).trim() || '🤔 No response from model',
+                text: String(text).trim() || 'No response from model',
                 usage: {
                     inputTokens: data.prompt_eval_count || 0,
                     outputTokens: data.eval_count || 0,
@@ -196,9 +239,10 @@ You are running locally on the user's machine. No data leaves their computer. Yo
             };
         }
         catch (error) {
+            console.error('[Agent] Ollama fetch error:', error);
             if (error.code === 'ECONNREFUSED' || error.message?.includes('fetch failed')) {
                 return {
-                    text: `⚠️ Ollama not running.\n\nPlease:\n1. Install Ollama: https://ollama.com\n2. Pull a model: ollama pull llama3.2\n3. Start Ollama: ollama serve`,
+                    text: `Ollama not running.\n\nPlease:\n1. Install Ollama: https://ollama.com\n2. Pull a model: ollama pull llama3.2\n3. Start Ollama: ollama serve`,
                 };
             }
             throw error;
@@ -206,8 +250,11 @@ You are running locally on the user's machine. No data leaves their computer. Yo
     }
     buildHistory(session, currentMessage) {
         const history = [];
-        // System prompt
-        const systemPrompt = session.systemPrompt || this.config?.systemPrompt || this.defaultSystemPrompt;
+        // System prompt with tools if enabled
+        let systemPrompt = session.systemPrompt || this.config?.systemPrompt || this.defaultSystemPrompt;
+        if (this.config?.enableTools && tools_1.toolRegistry.list().length > 0) {
+            systemPrompt += '\n\n' + (0, toolExecutor_1.getToolsSystemPrompt)();
+        }
         history.push({ role: 'system', content: systemPrompt, timestamp: 0 });
         // Previous messages
         for (const msg of session.messages) {
@@ -257,19 +304,19 @@ You are running locally on the user's machine. No data leaves their computer. Yo
     async processWithClaude(history) {
         // TODO: Implement Anthropic Claude API
         // Requires @anthropic-ai/sdk
-        return { text: '⚠️ Claude integration not yet implemented. Use Google/Gemini for now.' };
+        return { text: 'Claude integration not yet implemented. Use Google/Gemini for now.' };
     }
     async processWithOpenAI(history) {
         // TODO: Implement OpenAI API
         // Requires openai package
-        return { text: '⚠️ OpenAI integration not yet implemented. Use Google/Gemini for now.' };
+        return { text: 'OpenAI integration not yet implemented. Use Google/Gemini for now.' };
     }
     handleCommand(session, command) {
         const [cmd, ...args] = command.slice(1).split(' ');
         switch (cmd.toLowerCase()) {
             case 'status':
                 return {
-                    text: `📊 *Session Status*\n` +
+                    text: `*Session Status*\n` +
                         `Session ID: \`${session.id}\`\n` +
                         `Messages: ${session.messageCount}\n` +
                         `Model: ${session.model || this.config?.model || 'default'}\n` +
@@ -283,10 +330,10 @@ You are running locally on the user's machine. No data leaves their computer. Yo
                 const systemMsgs = session.messages.filter(m => m.role === 'system');
                 session.messages = systemMsgs;
                 session.messageCount = 0;
-                return { text: '🔄 Conversation reset. Let\'s start fresh!' };
+                return { text: 'Conversation reset. Let\'s start fresh!' };
             case 'help':
                 return {
-                    text: `📖 *ApexBot Commands*\n\n` +
+                    text: `*ApexBot Commands*\n\n` +
                         `/status - Show session info\n` +
                         `/new - Reset conversation\n` +
                         `/model <name> - Change AI model\n` +
@@ -296,11 +343,11 @@ You are running locally on the user's machine. No data leaves their computer. Yo
             case 'model':
                 if (args.length > 0) {
                     session.model = args.join(' ');
-                    return { text: `✅ Model changed to: ${session.model}` };
+                    return { text: `Model changed to: ${session.model}` };
                 }
                 return { text: `Current model: ${session.model || this.config?.model || 'default'}` };
             default:
-                return { text: `❓ Unknown command: /${cmd}\nType /help for available commands.` };
+                return { text: `Unknown command: /${cmd}\nType /help for available commands.` };
         }
     }
     getStatus() {