Fractalize Prompt is a tool that uses a dynamic, fractal network of agents instead of a
Sign in to like and favorite skills
Fractalize Prompt is a tool that uses a dynamic, fractal network of agents instead of a direct LLM or fixed network of agents. These dynamic agents recursively break down complex prompts into smaller tasks, which can be processed by a language model or other agents, and reassemble the results into a coherent and accurate response.
This system allows for more advanced and flexible prompt handling, ideal for tasks that require multi-step reasoning, context comprehension, and verification.
These requirements are extremely specific to my local computer and will be generalized in the future. They are also likely incomplete as my focus at the moment is on developing core functionality.
git clone https://gitlab.com/anadon/multiverse-school-summer-cohort-2-final-project-fractalizing-prompt-response cd fractalize-prompt
pip install -r requirements.txt
.envtouch .env
echo "ANTHROPIC_API_KEY=$ANTHROPIC_API_KEY" >> .env echo "OPENAI_API_KEY=$OPENAI_API_KEY" >> .env echo "WOLFRAM_APP_ID=$WOLFRAM_APP_ID" >> .env echo "BRAVE_SEARCH_API_KEY=$BRAVE_SEARCH_API_KEY" >> .env echo "LOCAL_LLAMA_API_KEY=$KEY_FROM_OLLAMA_SERVE" >> .env
poetry install poetry shell python -m fractalize_promptt
Testing is not implemented at this time. It doesn't make sense to implement until there is core functionality to test.
Please review the ROADMAP.md file for more information and contact me if interested.
Incomplete.
AGPLv3
A less restrictive license will be available later.
Josh (Anadon) Marshall