{"id":26608858,"url":"https://github.com/rajatasusual/llamabox","last_synced_at":"2026-04-11T12:04:23.503Z","repository":{"id":281421477,"uuid":"945227572","full_name":"rajatasusual/llamabox","owner":"rajatasusual","description":"A fully auto configured, self-hosted local AI \u0026 database stack on Debian WSL2.","archived":false,"fork":false,"pushed_at":"2025-03-23T14:11:59.000Z","size":1192,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"master","last_synced_at":"2025-03-23T15:24:16.648Z","etag":null,"topics":["ai","debian","edge-computing","llamacpp","localization","neo4j","rag","redis","self-hosted","slm","smolagents","wsl2"],"latest_commit_sha":null,"homepage":"","language":"Shell","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/rajatasusual.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2025-03-09T00:07:08.000Z","updated_at":"2025-03-23T14:12:42.000Z","dependencies_parsed_at":null,"dependency_job_id":"0543eb9f-c7f3-437a-922b-e5b2fdc5d66c","html_url":"https://github.com/rajatasusual/llamabox","commit_stats":null,"previous_names":["rajatasusual/wsl-assistant"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rajatasusual%2Fllamabox","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rajatasusual%2Fllamabox/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rajatasusual%2Fllamabox/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rajatasusual%2Fllamabox/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/rajatasusual","download_url":"https://codeload.github.com/rajatasusual/llamabox/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":245191600,"owners_count":20575247,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["ai","debian","edge-computing","llamacpp","localization","neo4j","rag","redis","self-hosted","slm","smolagents","wsl2"],"created_at":"2025-03-24T00:43:00.903Z","updated_at":"2026-04-11T12:04:23.474Z","avatar_url":"https://github.com/rajatasusual.png","language":"Shell","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Llamabox \u003cimg src=\"assets/icon.png\" width=\"32\" height=\"32\" style=\"vertical-align: middle\"\u003e\n\n[![Distro Health](https://github.com/rajatasusual/llamabox/actions/workflows/check.yml/badge.svg?branch=master)](https://github.com/rajatasusual/llamabox/actions/workflows/check.yml)  \n![WSL2](https://img.shields.io/badge/WSL2-Supported-blue) ![Debian](https://img.shields.io/badge/Debian-Supported-blue) ![License](https://img.shields.io/badge/License-MIT-green)\n\n\u003e **Run a full local RAG pipeline on your low-end, CPU-only Windows machine. Private, resilient, secure.**  \n\u003e **TL;DR:** See [SUMMARY.md](docs/SUMMARY.md) for the stack breakdown and performance.\n\n---\n\n## 🚀 What Is Llamabox?\n\n**Llamabox** lets you self-host a complete AI and database stack inside **WSL2 + Debian**, optimized for **CPU-only, low-resource machines**. It's built for privacy-first applications like chatbots, AI search, and offline assistants—**no GPU, no OpenAI keys, no cloud APIs**.\n\n🧠 Llamabox includes:\n- **llama.cpp**: CPU-only inference engine  \n- **Redis Stack**: Vector database for embeddings  \n- **Neo4j**: Graph-based knowledge base  \n- **Secure systemd-based setup**: Auto-restarting services  \n- **Optional browser extension**: Capture and sync content from Chrome/Edge\n\n![Memory usage screenshot](assets/mem.png)\n\n---\n\n## 📑 Table of Contents\n\n1. [Overview](#-what-is-llamabox)\n2. [Why WSL2 + Debian?](#-why-wsl2--debian)\n3. [Use Cases](#-use-cases)\n4. [Key Features](#-features)\n5. [Prerequisites](#-prerequisites)\n6. [Installation](#-installation)\n7. [Service Management](#-service-management)\n8. [Performance Benchmarks](#-performance-benchmarks)\n9. [FAQ \u0026 Troubleshooting](#-faq--troubleshooting)\n10. [Contributing](#-contributing)\n11. [License \u0026 Credits](#-license--credits)\n12. [Browser Extension](#-browser-extension)\n\n---\n\n## 🐧 Why WSL2 + Debian?\n\n- ✅ **WSL2** runs native Linux with low overhead on Windows  \n- ✅ **Debian** is lightweight and rock-stable  \n- ✅ **No Docker needed** – systemd and all services run directly under WSL  \n- ✅ Keeps everything local and private, with no cloud dependencies\n\n---\n\n## 🛠️ Use Cases\n\n- 💬 Local chatbots and AI assistants  \n- 🔍 Search over documents, pages, and structured data  \n- 🧩 Graph-based reasoning with Neo4j  \n- 🧠 Embed and store knowledge using Redis vectors  \n- 🛡️ Fully offline / air-gapped deployments\n\n---\n\n## 🧩 Features\n\n### 🛡️ Security\n- Passwordless local use; optional `fail2ban`, `ufw` for edge exposure\n- No SSH exposed by default\n- Auto security updates with `unattended-upgrades`\n\n### 🔄 Resilience\n- All critical services are systemd-managed\n- Auto-restarts on crash or reboot\n- Logs available via `journalctl`\n\n### 🧠 RAG Pipeline\n\u003e CPU-only, cloud-free, privacy-first Retrieval-Augmented Generation:\n1. User sends query → `llama.cpp`  \n2. Query embedding → Redis vector DB  \n3. Knowledge retrieved → Neo4j  \n4. Final answer generated → All local\n\n### ⚙️ Light on Resources\n- ~1GB idle memory usage\n- Runs on as little as 2 cores and 4GB RAM\n- Zero GPU required\n\n---\n\n## 🖥️ Prerequisites\n\n- Windows 10/11 with WSL2\n- Installed Debian distro via Microsoft Store or `wsl --install -d Debian`\n- Min. 4GB RAM (8GB recommended)\n- 20GB free disk space\n\n---\n\n## 📦 Installation\n\n```bash\n# In Windows Terminal:\nwsl --install -d Debian\n\n# Inside Debian WSL shell:\nsudo apt update \u0026\u0026 sudo apt install git -y\ngit clone https://github.com/rajatasusual/llamabox\ncd llamabox\n./setup.sh\n```\n\n🔧 See [INSTALLATION.md](docs/INSTALLATION.md) for customization and optional steps.\n\n---\n\n## 🔃 Service Management\n\n```bash\n# Check service statuses (Redis, Neo4j, llama-server, etc.)\n./scripts/check.sh\n\n# Or manually start a specific service:\nsudo systemctl restart neo4j\nsudo journalctl -u llama-server.service\n```\n\n📘 More in [MANAGE.md](docs/MANAGE.md)\n\n---\n\n## 🚀 Performance Benchmarks\n\n\u003e Test device: 4-core AMD Z1, 4GB RAM, WSL2 Debian  \n\u003e Model: **LLaMA 3B Q8_0**\n\n| Threads | Prompt Type | Tokens/sec | Notes |\n|---------|-------------|------------|-------|\n| 2       | `pp512`     | 253.07 ± 23.75 | Long-form |\n| 2       | `tg128`     | 54.44 ± 4.87   | Short query |\n\n![Tokens per second](assets/tpt.png)\n\n✅ Runs smoothly on CPU-only setup  \n✅ Great for background tasks and lightweight chatbots  \n✅ All on a 10-year-old laptop? Yes.\n\n---\n\n## ❗️ FAQ \u0026 Troubleshooting\n\n- ❓ **Systemd isn't working in WSL2**  \n  ✅ Add this to `/etc/wsl.conf`:\n  ```ini\n  [boot]\n  systemd=true\n  ```\n\n- ❓ **\"Out of memory\" loading model**  \n  ✅ Try a smaller GGUF model  \n  ✅ Or edit `.wslconfig` on Windows:\n  ```ini\n  [wsl2]\n  memory=8GB\n  ```\n\n- ❓ **Redis or Neo4j not starting?**  \n  ✅ Run `./scripts/check.sh`  \n  ✅ Or restart manually: `sudo systemctl restart redis-stack-server`\n\nMore in [FAQs.md](docs/FAQs.md)\n\n---\n\n## 🤝 Contributing\n\nWe’d love your help!  \n- Create issues, fix bugs, suggest features  \n- PRs welcome: fork → feature branch → pull request  \n- Style guide and guidelines coming soon\n\n---\n\n## 📄 License \u0026 Credits\n\nLicensed under the **MIT License**.  \nShout-outs:\n- [llama.cpp](https://github.com/ggml-org/llama.cpp)\n- [Redis Stack](https://redis.io/)\n- [Neo4j](https://neo4j.com/)\n- [WSL2](https://learn.microsoft.com/en-us/windows/wsl/) ❤️\n\n---\n\n## 🌐 Browser Extension\n\nThe [Llamabox Extension](https://github.com/rajatasusual/llamabox_extension) captures web pages and sends them to your local server for embedding.\n\n🔹 Features:\n- Extract full article text or selection\n- Sync with WSL2 HTTP server\n- Works offline, configurable shortcuts\n\n🔧 To install:\n- Clone the repo\n- Load `extension/` as an unpacked extension in Chrome or Edge\n- Set WSL IP in config page\n\nDocs: [llamabox_extension/README.md](https://github.com/rajatasusual/llamabox_extension/blob/master/README.md)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Frajatasusual%2Fllamabox","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Frajatasusual%2Fllamabox","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Frajatasusual%2Fllamabox/lists"}