Quantyval AI 🤖

Universal AI Agent Framework - Task execution, voice, multi-agent orchestration

Features

| Category | Features | |----------|----------| | Core | Agent with ReAct reasoning, LLM providers (OpenAI, Anthropic, OpenRouter, KiloCode) | | Security | Command allowlists, SSRF protection, API auth, rate limiting | | Memory | Short-term + long-term, Redis adapter for persistence | | Skills | Auto-detect tech stack (React, Node, Python, Go, etc.) | | CLI | Full command-line (run, serve, init, voice) | | Channels | Telegram bot, WebSocket, HTTP | | Tools | Playwright browser automation | | Multi-Agent | Sequential, parallel, hierarchical orchestration | | Voice | TTS (ElevenLabs, OpenAI) + STT (Whisper) |

Installation

npm install @quantyval-ai/quantyval-ai

Quick Start

CLI

# Run agent
npx quantyval run --model kilocode:kilo-auto/free

# Start server
npx quantyval serve --port 3000

# Initialize project
npx quantyval init

Code

import { Agent } from '@quantyval-ai/quantyval-ai';

const agent = new Agent({
  name: 'MyBot',
  systemPrompt: 'You are a helpful assistant.',
  llm: { type: 'kilocode', apiKey: process.env.QUANTYVAL_API_KEY },
});

const response = await agent.run('Hello!');
console.log(response.text);

Telegram Bot

TELEGRAM_BOT_TOKEN=xxx node examples/telegram-bot/bot.js

Configuration

| Option | Description | Default | |--------|-------------|---------| | name | Agent name | Quantyval | | model | LLM provider | kilocode:kilo-auto/free | | memory | Enable memory | false | | port | Server port | 3000 |

CLI Options

quantyval run --model openai:gpt-4        # Run interactive
quantyval serve --port 8080               # Start HTTP server
quantyval init                           # Create config
quantyval voice                          # Voice mode

Architecture

quantyval-ai/
├── bin/cli.js          # CLI entry point
├── src/
│   ├── core/           # Agent, LLM providers
│   ├── memory/         # Memory, Redis
│   ├── channels/       # Telegram, Discord
│   ├── tools/          # Browser automation
│   ├── server/         # HTTP server
│   ├── voice/          # TTS/STT
│   ├── multiagent/     # Orchestration
│   └── utils/          # Logger, errors, skills
├── examples/          # Example projects
└── tests/             # 37 tests

License

MIT