# RTGS2017/NagaAgent

A simple yet powerful agent framework for personal assistants, designed to enable intelligent interaction, multi-agent collaboration, and seamless tool integration.

Repository: https://github.com/RTGS2017/NagaAgent
Canonical: https://ross.abutalabs.com/products/nagaagent
Homepage: https://space.bilibili.com/266938091
Language: Python
License: AGPL-3.0
License Family: copyleft
Topics: agentic-ai, agents, mcp-client, openai, mcp, toolcall, openclaw
Last push: 2026-08-11T07:55:16+00:00

## Health v2 (maintenance only)
Score: 83/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 97, release rhythm 93, longevity 35
- inputs: {"age_days": 501, "days_push": 22, "days_rel": 49, "gap_med": 7, "n_releases_24m": 6}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1544, forks 167 (observed 2026-08-28T04:05:01.384127+00:00)

## What it is
NagaAgent is a desktop AI personal assistant framework with an anime (Live2D) virtual character, built in Python with an Electron front end. It combines streaming tool calls, MCP client support, knowledge-graph-based memory, voice interaction (ASR/TTS), and an OpenClaw-powered web exploration mode into a consumer-facing assistant application.

## Use cases
- run an anime-style AI assistant on my desktop with a Live2D avatar
- chat with an AI assistant that remembers past conversations via a knowledge graph
- let an AI agent automatically explore the web and complete tasks for me
- connect my assistant to tools through MCP servers
- get voice conversations with a personal AI assistant using speech recognition and TTS
- use an AI assistant that can read the game screen and suggest strategies
- build a personal assistant that calls tools and multiple agents collaborate

## When to choose
- you want a ready-to-use desktop AI companion with a Live2D character, voice, and memory rather than a bare library
- you need MCP client support and streaming tool calls in a personal assistant
- you want automatic web exploration (OpenClaw) and browser control built into the assistant
- you want a game-assistant mode that can read game screens and integrate automation plugins like MAA

## When to avoid
- you need a minimal headless agent framework to embed in your own backend service
- you require a permissive license - the core is AGPL-3.0 with proprietary dual-licensing for closed use
- you need first-class Linux server or mobile deployment - it targets desktop platforms
- you want a fully English-first product - documentation and UX are Chinese-centric

## Facets
- artifact type: application
- maturity: active
- function: agent-framework, mcp, chatbot, tts, speech-recognition, llm-inference, web-scraping, gui, plugin-system
- domain: large-language-models, chatbots, artificial-intelligence, desktop-applications, gaming-tools, cross-platform
- platform: windows, python
- tags: live2d, virtual-assistant, multi-agent, tool-calling, knowledge-graph-memory, voice-interaction, anime-assistant, openclaw, game-assistant, personal-assistant, ai-agents, automation, macos, linux, desktop, electron

## Member repositories
- RTGS2017/NagaAgent (main) score 83

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:01.384127+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:30:37.066014+00:00, confidence not recorded.
  - readme: https://github.com/RTGS2017/NagaAgent (fetched 2026-08-28T04:05:01.384127+00:00, sha a0e7b35d037c)
- Data as of 2026-08-30T08:39:29.467469+00:00.