This project implements an AI-powered brainstorming system using the CrewAI framework. It creates a team of specialized AI agents that work together to generate, analyze, and evaluate ideas on a given topic.

• Team of specialized AI agents:
  • Creative Idea Generator
  • Practical Analyst
  • UX Expert
  • Technical Expert
  • Strategic Evaluator
• Structured brainstorming process
• Token usage tracking
• Ready-made templates for common topics
• Detailed reports in both JSON and readable formats
• Support for Venice AI models

• Python 3.8 or higher
• Venice AI API key

1. Clone the repository:

git clone https://github.com/venticello/venice-brainstorm.git
cd venice-brainstorm

2. Install dependencies:

pip install -r requirements.txt

3. Create a .env file in the project root with your Venice AI API key:

VENICE_API_KEY=your_api_key_here

Run the script with a default template:

cd src
python brainstorm.py

Run with a custom topic:

cd src
python brainstorm.py --topic "Your custom topic" --context "Additional context"

• --topic: Custom topic for brainstorm
• --context: Additional context for the topic
• --template: Template to use (product, ecology, education, startup)
• --temperature: Temperature parameter for LLM (0.0 to 1.0)
• --top-p: Top-p parameter for LLM (0.0 to 1.0)
• --model: Model to use (list bellow)

• venice-uncensored 32768
• qwen-2.5-qwq-32b 32768
• qwen3-4b 32768
• mistral-31-24b 131072
• qwen3-235b 32768
• llama-3.2-3b 131072
• llama-3.3-70b 65536
• llama-3.1-405b 65536
• dolphin-2.9.2-qwen2-72b 32768
• qwen-2.5-vl 32768
• deepseek-r1-671b 131072

You can configure the following settings in the .env file:

• VENICE_API_KEY: Your Venice AI API key
• BRAINSTORM_MODEL: AI model to use
• VENICE_TEMPERATURE: Default temperature setting
• VENICE_TOP_P: Default top-p setting

The script generates two types of output files in the results directory:

1. JSON report (venice_brainstorm_TIMESTAMP.json)
2. Readable text report (venice_brainstorm_TIMESTAMP_readable.txt)

The reports include:

• Topic and context
• Results from each agent
• Final evaluation and recommendations
• Token usage statistics

.
├── src/
│   ├── brainstorm.py    # Main script
│   ├── agents.py        # Agent definitions
│   ├── tasks.py         # Task definitions
│   ├── templates.py     # Ready-made templates
│   ├── config.py        # Configuration settings
│   └── utils.py         # Utility functions
├── results/             # Directory for output files
├── requirements.txt     # Project dependencies
└── README.md           # This file

cd src
python brainstorm.py --template startup

This will run a brainstorming session on creating an AI technology startup with a budget of $100k and a 12-month planning horizon.

MIT License