ShaneLiu04

ShaneLiu04 / NeoAgent

Public

支持多 Agent 编排、流式透明与安全沙箱的 AI Agent 框架。

14
0
85% credibility
Found May 27, 2026 at 14 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

NeoAgent is an open-source Python framework for building AI assistants that can handle multiple tasks simultaneously, remember user preferences, and safely execute code. It features a colorful terminal interface with an animated companion character, real-time streaming of responses, multi-agent parallel task execution, a three-layer memory system (short-term conversation memory, long-term vector-based memory, and structured user profiles), cost-aware routing that automatically switches between affordable and powerful AI models, Docker-based code execution sandboxing, web search integration, task scheduling with desktop notifications, and image analysis for supported models. The framework includes production-grade observability for tracking costs and usage, supports multiple AI providers (DeepSeek, OpenAI, Anthropic, Google, Groq, Ollama), and provides both a command-line interface and a web API server.

How It Works

1
🔍 You discover NeoAgent

You find an open-source AI assistant framework that promises to handle multiple tasks at once, remember your preferences, and run code safely.

2
📦 You install it on your computer

With one command, the framework installs alongside its colorful terminal interface.

3
🔑 You connect your AI service

You enter your API key for your preferred AI provider (like DeepSeek, OpenAI, or Anthropic) so the assistant can think and respond.

4
🤖 You meet N.E.O., your cyberpunk companion

A friendly animated character appears in your terminal and reacts to everything the assistant does—celebrating when tasks complete, warning you during errors.

5
💬 You ask questions naturally

You type questions like 'compare Python, Rust, and Go for web development' and watch tokens stream in real-time as the assistant thinks.

6
The assistant works its magic
🔬
Parallel research mode

It studies multiple topics simultaneously, then synthesizes everything into one clear report.

💻
Code execution mode

It runs Python code in a secure sandbox to verify answers or run calculations.

🔔
Reminder mode

You set a reminder for later, and your computer chimes with a desktop notification when it's time.

7
🧠 It remembers you

The assistant stores your preferences and facts about you, so next time it can personalize responses automatically.

🎉 You get your results

Whether it's a comprehensive comparison, working code, or a timely reminder, everything works exactly as promised—and you see exactly what it's doing every step of the way.

Sign up to see the full architecture

6 more

Sign Up Free

Star Growth

See how this repo grew from 14 to 14 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is NeoAgent?

NeoAgent is a Python framework for building AI agents with multi-agent orchestration, real-time streaming, and secure code execution. It lets you spawn parallel sub-agents for independent tasks, route requests between cheap and expensive models based on query complexity, and run Python code in Docker containers without risking your system. The framework provides three-layer memory (recent conversation context, vector-based long-term storage via ChromaDB, and structured user profiles), per-session cost tracking, and an interactive terminal interface with a cyberpunk companion character named "N.E.O." that reacts to agent activity. It builds on LangGraph for state machine orchestration and supports DeepSeek, OpenAI, Anthropic, Google, Groq, and Ollama through a unified provider factory.

Why is it gaining traction?

The standout features are its multi-agent system that lets you say "research A, B, and C in parallel" in natural language, automatic cost optimization by routing simple queries to cheap models, and real OS-level sandboxing via Docker rather than regex-based path blocking. Every token, tool call, and budget decision emits a typed event you can intercept in real time. The observability layer tracks costs, latency, and token usage out of the box. Compared to building these pieces separately or using less mature frameworks, NeoAgent bundles the essential production requirements into a coherent package.

Who should use this?

Teams building agent pipelines that need parallel task execution across specialized agents. Developers running untrusted code who need proper container isolation. Python shops running high-volume agent applications where cost routing could meaningfully reduce API bills. Developers who want LangGraph orchestration but need better defaults for streaming, security, and cost control than rolling their own.

Verdict

The credibility score of 0.85% reflects an extremely new project with only 14 stars and minimal community validation. The feature set is solid and the 65-test suite covers real scenarios well, but adopting this means betting on an unproven codebase. Worth exploring for experimentation or internal tooling where community support gaps are acceptable, but production deployments should wait for stronger signals of maintenance and adoption.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.