NOVA THE LOCAL AI ASSISTANT
Описание и применение
NOVA – локальный AI‑ассистент для Windows, объединяющий голосовое взаимодействие, локальные модели, память и автоматизацию, снижая повторяющиеся задачи.
Задача
Users repeatedly switch applications, search files, and repeat instructions, leading to inefficiency in daily digital work.
Результат
NOVA consolidates repetitive digital tasks into a single local AI assistant, enabling offline voice interaction, persistent memory, permission‑controlled desktop automation, and visual context awareness, thereby reducing manual effort and improving workflow efficiency.
Рабочий процесс
- 1. Capture voice input via sounddevice and openWakeWord for wake‑word detection.
- 2. Transcribe audio with Vosk offline speech recognition.
- 3. Generate response using local Ollama runtime with Qwen 2.5 models.
- 4. Synthesize speech with Piper TTS for voice output.
- 5. Persist conversation, project, and task data in SQLite memory using FastEmbed for semantic retrieval.
- 6. Execute desktop automation commands through registered tools with permission checks, previews, and logging.
- 7. Capture screen and perform OCR for visual context while filtering sensitive data.
- 8. Coordinate tasks across Research, Developer, System, Vision, and Memory agents via a workflow orchestrator.
Доказательства
I built NOVA (Neural Operational Virtual Assistant), a local-first, JARVIS-inspired AI desktop assistant designed to take care of repetitive computer tasks and make everyday digital work easier.
It supports: Local AI conversations, Offline voice interaction, Persistent memory, Desktop automation, Visual intelligence, Multi-agent orchestration, Proactive automation, Personalization.
NOVA is built around open-source technologies and open-weight AI models. My goal was to avoid making the assistant dependent on a single closed AI provider.
I use Ollama as the local model runtime, with Qwen 2.5 models for language generation.
NOVA includes a registered desktop tool system with 23 tools covering application, file, and system operations.
- Дата записи: 2026-10-26T05:26:18.315Z