llm-brand-generator

An AI-powered branding assistant that generates brand name, slogan, and concept visuals from a simple description.
It uses LangChain + Ollama for text generation and (for now) a simple placeholder image generator.

---

How to use it?

Simply provide a description of your desired brand AND click on 'Generate Branding':

After approximately one minute, you will receive the recommended items:

• A suggested name
• A suggested slogan
• A branding concept
• A logo mark
• A palette of colors

---

---

How its works?

This project combines local LLMs with cloud APIs to generate branding suggestions:

1. Ollama (local LLM runtime)

   • You need to have Ollama installed on your system.
   • The Python code (via langchain-community) connects to Ollama’s local server (default: http://localhost:11434) using the model llama3.2.
   • Ollama handles downloading, storing, and running the model.

2. LangChain (retrieval + prompting)

   • Loads reference documents from the docs/ folder (PDF/TXT).
   • Uses FAISS to build or reuse a vector store.
   • Provides context to the LLM for better branding suggestions.

3. Branding generation process

   • You describe your business/idea in the Streamlit interface.
   • The app queries Ollama with context + your description.
   • The model outputs:
     • Brand Name
     • Slogan
     • Logo Mark idea
     • Color
     • Branding Concept

4. Visual generation

   • Logo image is generated using Pollinations.ai.
   • Color palette is fetched from ColorMagic and overlaid with the slogan.

System Architecture

flowchart TD
    User([User Input]) --> Streamlit[Streamlit UI]
    Streamlit --> Agent[LangChain Agent]
    Agent -->|RAG| Vectorstore[(FAISS Vectorstore)]
    Agent --> Ollama((Ollama llama3.2))
    Agent --> Pollinations[Pollinations.ai - Logo Image]
    Agent --> Colormagic[ColorMagic API - Palettes]
    Ollama --> Result[Branding Results]
    Pollinations --> Result
    Colormagic --> Result
    Result --> Streamlit

Loading

---

Setup

1. Install and run Ollama

• Make sure you have Ollama installed on your machine.
• If not, download it here: https://ollama.com/download
• Then pull the required model:

ollama pull llama3.2

You can test if the model is working by running:

ollama run llama3.2 "Hello world"

2. Clone the repository

git clone https://github.com/gil-son/llm-brand-generator.git

3. Create a virtual environment

cd llm-brand-generator/
python3 -m venv venv

4. Activate the environment

source venv/bin/activate

5. Install dependencies

pip install -r requirements.txt

streamlit run app.py

6. Run the app

streamlit run app.py

Debug

• Basic way: just run this command

streamlit run app.py --logger.level=debug

• Step by step (VS Code): create a .vscode/ folder and inside create a launch.json file with:

{
  "version": "0.2.0",
  "configurations": [
    {
      "name": "Debug Streamlit",
      "type": "python",
      "request": "launch",
      "program": "app.py",
      "args": [
        "run",
        "app.py",
        "--logger.level=debug"
      ],
      "console": "integratedTerminal"
    }
  ]
}

For conditional parts, just use something trivial like if 2 > 1 to ensure the code executes and logging continues.

• Interactive debug (terminal only): this will pause execution in the terminal, and you can inspect variables step by step.

import pdb; pdb.set_trace()

Project Structure

llm-brand-generator/
├─ app.py                # main interface (Streamlit + integration)
├─ agent.py              # LangChain agent + RAG pipeline
├─ assets.py             # base64 images and icons for the interface
├─ pollinations_api.py   # image generation (Pollinations.ai)
├─ colormagic_api.py     # integration with ColorMagic API (color palettes)
├─ embeddings/           # scripts to generate embeddings
│   └─ build_embeddings.py
│
├─ docs/                 # reference documents (PDF, TXT)
│   ├─ branding_guide.txt
│   └─ marketing_tips.pdf
│
├─ requirements.txt      # project dependencies
└─ vectorstore/          # FAISS store (created automatically)
   └─ index.faiss

---