Project N.E.K.O.Developer Documentation

A proactive, omni-modal AI companion featuring 24/7 ambient awareness, agent capability and an embodied emotional engine.

Get Started

Get on Steam

API Reference

View on GitHub

🎮

Steam Workshop & Community

Available on Steam with full Workshop UGC support — share and discover characters, models, voice packs, and plugins created by the community.

View on Steam

🎙️

Omni-Modal Dialogue

Voice, text, and vision in a unified conversation loop. Real-time speech with RNNoise neural denoising, AGC, and VAD for ultra-low-latency interaction.

Learn more

💬

Proactive Chat

24/7 ambient awareness — screen understanding, social media trends, personal feeds, music & memes. She initiates conversations with you.

Learn more

🧠

Three-Tier Memory

Semantic recall via hybrid embedding-vector and BM25 indexing. Facts, reflections, and persona layers with sliding-window compression and persistent user preferences.

How it works

🤖

Agent Framework

Background task execution via MCP tools, Computer Use, Browser Use, and OpenFang A2A adapters. Automatic task planning, deduplication, and parallel execution.

Explore agents

🔌

Plugin Ecosystem

Full plugin SDK & marketplace for custom extensions. Decorator-based API, async lifecycle hooks, and inter-plugin messaging. Built-in plugins for MCP, reminders, livestreaming, smart home, and more.

Build a plugin

🎭

Live2D, VRM, MMD & Voice Clone

Embodied avatars with emotion-mapped expressions, lip sync, and idle animations. Clone any voice from a 5-second sample via MiniMax or CosyVoice backends.

Frontend guide

🌐

14+ AI Providers & i18n

OpenAI, Anthropic, Google, Qwen, DeepSeek, Groq, Ollama, and more — with free models out of the box. Full UI and prompt localization across 6 languages (zh-CN, zh-TW, en, ja, ko, ru).

Provider list