Lite-Brain

Lite-Brain is a lightweight, fast, and high‑performance application designed to run on mid‑range to high‑range computers. It allows users to have their own local artificial brain, with a clean interface that covers 90% of typical everyday needs.

---

🚀 Setup

Click on the launcher LiteBrainLauncher.exe. Do your configuration and start !

---

🎯 Goals

• Reduce dependency on large LLMs (ChatGPT, Claude, etc.) for standard tasks. → Lower environmental cost, fully local processing, no data leakage.

• Runs smoothly on a PC with a small RTX GPU.

• Three modes available from the home screen or a dropdown menu:

  • 🌿 Classic Mode (recommended) --- fast and lightweight
  • ⚡ Turbo Mode --- more accurate but more resource‑intensive
  • ⚡⚡ Ultra Mode --- highly accurate, requires a strong machine

---

🧠 Features

1. Writing & Rewriting Assistance

• Write emails, messages, letters, articles, reports.
• Rewrite, simplify, or enhance text.
• Grammar/spelling/style correction.
• Generate summaries (text, PDF, transcripts).

2. Document Synthesis & Analysis

• Extract key information from files, images, notes, PDFs.
• Summaries, key points, digests.
• Search within your local documents.
• Q&A based on local files (PDF, DOCX, Markdown, TXT...).

3. Brainstorming & Idea Generation

• Project ideas, posts, titles, slogans.
• Arguments and counter‑arguments.
• Create plans and structured steps.

4. Programming Assistance

• Explain code.
• Fix or complete simple functions.
• Generate small scripts (Python, JS, Bash...).
• Create documentation or comments.

5. Learning & Concept Explanation

• Simplify complex concepts.
• Provide examples and analogies.
• Create quizzes or exercises.

6. Translation & Style Adaptation

• Fast language translation.
• Rewriting for tone, audience, level.
• Adapt content to technical, marketing, etc.

7. Organization & Planning

• Generate action plans, calendars, to‑do lists.
• Structure projects or summaries.
• Brainstorm outlines for articles or training content.

8. Conversational Assistant

• A fully local chat assistant.
• General advice (productivity, habits, organization...).

---

🛠️ Tech Stack

• UI: FastAPI
  • Front: Jinja / HTMX / Tailwind
  • Backend + SQLite (light mono‑user DB)
  • File upload and RAG systems.
• ML: FastAPI
  • Embeddings bge‑m3
  • OCR
  • Reranking bge-v2-v3
  • Transcription whisper TO DO LATER
• Ollama: LLM server
  • Qwen3:4B for Classic Mode
  • Qwen3:8B for Turbo Mode
  • Qwen3:14B for Ultra Mode
• Qdrant: Vector database
  • Command for Docker Local run:

    docker run --name temp_qdrant -p 6333:6333 -v qdrant_storage:/qdrant/storage qdrant/qdrant:v1.15

---

Build the Launcher

For building the LiteBrainLauncher.exe i use pyinstaller and run the command :

🔹 Windows : pyinstaller launcher.py --onefile --name "LiteBrainLauncher" --icon logo/lite-brain-logo.ico

🔹 MacOS : pyinstaller launcher.py --onefile --name "LiteBrainLauncher" --icon logo/lite-brain-logo.icns

🔹 Linux : pyinstaller launcher.py --onefile --name "LiteBrainLauncher" --icon logo/lite-brain-logo.png

---

🧩 Architecture Schema

---

Author

• @nixiz0