GN

gofenix/nex-agent

Developer tools

이 페이지는 AI로 자동 번역되었습니다. 기술적인 세부 사항은 원문을 확인하세요. 원문 보기

70 stars 품질 48 트렌드 22

Your own self-evolving AI agent powered by Elixir/OTP

개요

NexAgent Your own self-evolving AI agent A long-running agent that works in the chat apps you already use, calls tools, remembers context, and keeps improving through real-world use. 中文文档 is an AI agent built for real-world, long-running use. It is not just a one-shot CLI demo, and it is not just a thin prompt wrapper around a model. NexAgent is built around a more specific goal: keep an agent online, place it inside the chat apps you already use, give it memory and tools, let it manage background work, and make it capable of improving over time. - : not only prompt engineering, but also memory, skills, tools, and source-level self-improvement. - : supervision trees, GenServers, process isolation, and hot code loading for fault tolerance, concurrency, and long-lived operation.

README

원문 README의 일부입니다. 원문 보기.

NexAgent is an AI agent built for real-world, long-running use.

It is not just a one-shot CLI demo, and it is not just a thin prompt wrapper around a model. NexAgent is built around a more specific goal: keep an agent online, place it inside the chat apps you already use, give it memory and tools, let it manage background work, and make it capable of improving over time.

Two ideas define the project today:

  • Self-evolution: not only prompt engineering, but also memory, skills, tools, and source-level self-improvement.
  • Elixir/OTP: supervision trees, GenServers, process isolation, and hot code loading for fault tolerance, concurrency, and long-lived operation.

At a Glance

If you only remember three things about NexAgent, they should be:

  • What it is: a long-running AI agent that lives in chat apps
  • What it can do: memory, tools, skills, scheduled jobs, and background tasks
  • Why it can keep running: Elixir/OTP plus built-in self-evolution paths
flowchart TD
    A["Chat AppsTelegram / Feishu / Discord / Slack / DingTalk"] --> B["Gateway"]
    B --> C["InboundWorker"]
    C --> D["Runner"]
    D --> E["Sessions + Memory"]
    D --> F["Tools + Skills"]
    D --> G["Cron + Subagent"]
    D --> H["Reflect + Upgrade Code"]
    H --> I["CodeUpgrade + UpgradeManager"]

What You Can Build

Use case What NexAgent is doing
Always-on assistant Stays in your chat apps and keeps per-chat context over time
Personal knowledge agent Combines long-term memory, history, and retrieval
Automation assistant Uses cron for reminders, recurring jobs, and background work
A growing agent Expands through skills, tools, and code-level self-improvement

Key Features

Capability What it means
Self-evolving by design soul_update, memory_write, skill_capture, tool_create, reflect, and upgrade_code keep the agent from staying static
Long-running sessions Sessions are scoped by channel:chat_id and keep memory, history, and isolation
Works in your chat apps Telegram, Feishu, Discord, Slack, and DingTalk are already supported
Tools, skills, and memory built in File access, shell, web, messaging, memory search, scheduling, and skills come built in
Background work included Cron jobs and subagents are part of the system, not an afterthought
Built on Elixir/OTP Supervision trees, service processes, and hot reload support real-world uptime

Why NexAgent

GitHub에서 전체 README 보기

추천 도구

다른 키워드를 입력하거나 필터를 제거해 보세요.

설치

npx skillfish add gofenix/nex-agent
MCP 초보자라면? mcpskill.org 가이드부터 시작하세요 →