AI Skill Hub 强烈推荐:server-nexe AI技能包 是一款优质的AI工具。AI 综合评分 8.0 分,在同类工具中表现稳健。如果你正在寻找可靠的AI工具解决方案,这是一个值得深入了解的选择。
支持持久内存、RAG和多后端推理的本地AI服务器
server-nexe AI技能包 是一款基于 Python 开发的开源工具,专注于 ai、apple-silicon、embeddings 等核心功能。作为 GitHub 开源项目,它拥有活跃的社区支持和持续的版本迭代,代码完全透明可审计,支持本地部署以保护数据隐私。无论是个人使用还是集成到企业工作流,都能提供稳定可靠的解决方案。
支持持久内存、RAG和多后端推理的本地AI服务器
server-nexe AI技能包 是一款基于 Python 开发的开源工具,专注于 ai、apple-silicon、embeddings 等核心功能。作为 GitHub 开源项目,它拥有活跃的社区支持和持续的版本迭代,代码完全透明可审计,支持本地部署以保护数据隐私。无论是个人使用还是集成到企业工作流,都能提供稳定可靠的解决方案。
# 方式一:pip 安装(推荐)
pip install server-nexe
# 方式二:虚拟环境安装(推荐生产环境)
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install server-nexe
# 方式三:从源码安装(获取最新功能)
git clone https://github.com/jgoy-labs/server-nexe
cd server-nexe
pip install -e .
# 验证安装
python -c "import server_nexe; print('安装成功')"
# 命令行使用
server-nexe --help
# 基本用法
server-nexe input_file -o output_file
# Python 代码中调用
import server_nexe
# 示例
result = server_nexe.process("input")
print(result)
# server-nexe 配置文件示例(config.yml) app: name: "server-nexe" debug: false log_level: "INFO" # 运行时指定配置文件 server-nexe --config config.yml # 或通过环境变量配置 export SERVER_NEXE_API_KEY="your-key" export SERVER_NEXE_OUTPUT_DIR="./output"
<p align="center"> <img src=".github/logo.svg" alt="server.nexe" width="400"> </p>
<p align="center"> <strong>Local AI server with persistent memory. Zero cloud. Full control.</strong> </p>
<p align="center"> <em>I've reached the minimum viable product for the real world — but feedback is still missing. 🚀</em> </p>
<p align="center"> <a href="https://github.com/jgoy-labs/server-nexe/actions/workflows/ci.yml"><img src="https://github.com/jgoy-labs/server-nexe/actions/workflows/ci.yml/badge.svg" alt="CI"></a> <img src=".github/badges/coverage.svg" alt="Coverage"> <a href="LICENSE"><img src="https://img.shields.io/badge/license-Apache%202.0-blue" alt="License"></a> <a href="https://www.python.org"><img src="https://img.shields.io/badge/python-3.11%2B-blue?logo=python&logoColor=white" alt="Python"></a> <a href="https://fastapi.tiangolo.com"><img src="https://img.shields.io/badge/FastAPI-0.136-009688?logo=fastapi&logoColor=white" alt="FastAPI"></a> <a href="https://v2.tauri.app"><img src="https://img.shields.io/badge/Tauri%20v2-desktop%20app-FFC131?logo=tauri&logoColor=white" alt="Tauri v2"></a> </p>
<p align="center"> <a href="https://qdrant.tech"><img src="https://img.shields.io/badge/Qdrant-vector--db-dc244c?logo=qdrant&logoColor=white" alt="Qdrant"></a> <a href="https://github.com/ml-explore/mlx"><img src="https://img.shields.io/badge/MLX-Apple%20Silicon-000000?logo=apple&logoColor=white" alt="MLX"></a> <a href="https://ollama.com"><img src="https://img.shields.io/badge/Ollama-compatible-black?logo=ollama&logoColor=white" alt="Ollama"></a> <a href="https://github.com/ggerganov/llama.cpp"><img src="https://img.shields.io/badge/llama.cpp-GGUF-8B5CF6" alt="llama.cpp"></a> <a href="https://github.com/jgoy-labs/server-nexe"><img src="https://img.shields.io/badge/RAG-local%20%7C%20private-22c55e" alt="RAG"></a> <a href="https://github.com/sponsors/jgoy-labs"><img src="https://img.shields.io/badge/sponsor-♥-ea4aaa?logo=github-sponsors&logoColor=white" alt="Sponsor"></a> </p>
<p align="center"> <a href="https://server-nexe.org"><strong>Documentation</strong></a> · <a href="#-quick-start"><strong>Install</strong></a> · <a href="#-architecture"><strong>Architecture</strong></a> · <a href="https://github.com/jgoy-labs/server-nexe/releases"><strong>Releases</strong></a> </p>
<p align="center"> <a href="README-ca.md"><strong>Català</strong></a> · <a href="README-es.md"><strong>Español</strong></a> </p>
---
v1.0.7 — Memory & collections fixes, Windows ARM64 support. Server Nexe now ships as a Tauri v2 desktop application with onboarding wizard, system tray, and automatic sidecar management. Available as macOS DMG (Apple Silicon), Linux AppImage (ARM64), and Windows ARM64 installer (unsigned — SmartScreen warns). See Releases. Linux note: tested on Ubuntu 24.04 ARM64 virtual machines (UTM). CPU inference (Ollama) verified. If you test on native hardware or with GPU acceleration, please open an issue with your results. Windows note: Windows ARM64 is supported since v1.0.7. The NSIS installer is unsigned — SmartScreen will warn: choose "More info" → "Run anyway". The installer handles WebView2; the app then installs Ollama (the inference engine on Windows) automatically on first run, from the onboarding wizard.
---
RAM and disk are the same across platforms (they depend on the model you pick, not the OS). OS, CPU/arch and the available inference backend differ:
| macOS | Linux | Windows | |
|---|---|---|---|
| **OS** | 14 Sonoma+ | Ubuntu 24.04+ (tested on VM) | 11 ARM64 (since v1.0.7) |
| **CPU / arch** | Apple Silicon (M1+) | x86_64 or ARM64 | ARM64 |
| **Inference backend** | MLX + llama.cpp + Ollama | Ollama (+ llama.cpp on x86_64) | Ollama only |
| **Extra native deps** | — (bundled) | WebKitGTK 4.1 | WebView2 (installer handles it) |
| Common (all platforms) | Minimum | Recommended |
|---|---|---|
| **RAM** | 8 GB | 16 GB+ (for larger models) |
| **Disk** | 10 GB free | 20 GB+ free |
| **Python** | 3.11+ (only needed when installing from source; the installer bundles it) | 3.12+ |
macOS: Apple Silicon only (arm64). Intel Macs and macOS 13 Ventura are no longer supported since v0.9.9. Backends caveat: models marked "MLX Apple Silicon" in the model catalog run only on macOS. On Windows and Linux ARM64 (Ollama-only) those models are unavailable — pick an Ollama-backed model.
NEXE_AUTOSTART_OLLAMA=true pytest -m "integration" -q ```
Web UI — light mode |
Web UI — dark mode |
Download the latest package from Releases:
| Platform | Package | Size |
|---|---|---|
| macOS (Apple Silicon) | nexe-app_1.0.7_aarch64.dmg | ~1.3 GB |
| Linux (ARM64) | nexe-app_1.0.7_aarch64.AppImage | ~1.2 GB |
| Windows (ARM64) | nexe-app_1.0.7_arm64-setup.exe (unsigned — SmartScreen warns) | ~1.3 GB |
The onboarding wizard handles everything: hardware detection, backend selection, model download, and configuration. The app runs server-nexe as a sidecar process with system tray integration.
git clone https://github.com/jgoy-labs/server-nexe.git
cd server-nexe
./setup.sh # guided installation (detects hardware, picks backend & model)
nexe go # start server on port 9119
Once running:
nexe chat # interactive chat (RAG memory on by default)
nexe memory store "Barcelona is the capital of Catalonia"
nexe memory recall "capital Catalonia"
nexe status # system status
```bash
echo '{"model_key": "qwen35_4b", "engine": "ollama"}' | python -m installer.install_headless nexe go ```
Endpoints at http://localhost:9119:
| Endpoint | Description |
|---|---|
/v1/chat/completions | OpenAI-compatible chat API |
/ui | Web UI (chat, file upload, sessions) |
/health | Health check |
/docs | Interactive API documentation (Swagger) |
Authentication viaX-API-Keyheader. Key is generated during installation and stored in.env.
Auto-discovered plugins with independent manifests. Security, web UI, RAG, backends — everything is a plugin. Add capabilities without touching the core. NexeModule protocol with duck typing, no inheritance.
</td> <td width="50%">
Server Nexe uses a duck typing protocol (NexeModule Protocol) — no class inheritance, no BasePlugin. Each plugin is a directory under plugins/ with a manifest.toml and a module.py.
5 active plugins:
| Plugin | Type | Key features |
|---|---|---|
| **mlx_module** | LLM Backend | Apple Silicon native, prefix caching (trie), Metal GPU |
| **llama_cpp_module** | LLM Backend | Universal GGUF, LRU ModelPool, CPU/GPU |
| **ollama_module** | LLM Backend | HTTP bridge to Ollama, auto-start, VRAM cleanup |
| **security** | Core | Dual-key auth, 6 injection detectors + NFKC, 49 jailbreak patterns, rate limiting, RFC5424 audit logging |
| **web_ui_module** | Interface | Web chat, sessions, document upload, MEM_SAVE, RAG sanitization, i18n |
高性能本地AI服务器,支持多种后端推理
该工具使用 NOASSERTION 协议,商用场景请仔细阅读协议条款,必要时咨询法律意见。
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
📄 NOASSERTION — 请查阅原始协议条款了解具体使用限制。
总体来看,server-nexe AI技能包 是一款质量优秀的AI工具,在同类工具中具备一定竞争力。AI Skill Hub 将持续追踪其更新动态,建议收藏备用,结合自身场景选择合适时机引入使用。
| 原始名称 | server-nexe |
| 原始描述 | 开源AI工具:Local AI server with persistent memory, RAG, and multi-backend inference (MLX / 。⭐9 · Python |
| Topics | aiapple-siliconembeddingsfastapillama-cpppython |
| GitHub | https://github.com/jgoy-labs/server-nexe |
| License | NOASSERTION |
| 语言 | Python |
收录时间:2026-05-16 · 更新时间:2026-05-30 · License:NOASSERTION · AI Skill Hub 不对第三方内容的准确性作法律背书。