OllarooGPT - built with love and passion

It's a lightweight chat assistant integrated with the Ollama Mistral model. Its functionalities include:

• Handling prompts and generating responses.
• Accepting photos and providing descriptions.
• Generating songs from text.
• Implementing Retrieval-Augmented Generation (RAG): Upload a PDF, and the LLM provides responses based on the prompt, with vector data chunked and stored in Pinecone.

Upcoming additions will include integration with

• OpenAI's DALL·E
• Reasoning APIs.
• Document selection where individual pdf is store separately on different collections or index.

Watch this YouTube video

Building docker image

docker build --no-cache -t kenken64/chatgpt --build-arg PINECONE_API_KEY=xxx --build-arg PINECONE_INDEX_NAME=pdf-genai --build-arg SUNO_API_URL=http://host.docker.internal:3001/api/ --build-arg OLLAMA_BASE_URL=http://host.docker.internal:11434 --build-arg OLLAMA_MODEL=llama3.1:8b --build-arg VISUAL_MODEL=minicpm-v --build-arg CORBADO_PROJECTID=xxxx --build-arg CORBADO_API_KEY=xxxx --build-arg APP_PORT=3000 .

Running docker image

--rm: This flag tells Docker to remove the container when the process running inside it exits.

docker run -p 3000:3000 --rm -d --name kenken64-chatgpt -e PINECONE_API_KEY=xxx -e PINECONE_INDEX_NAME=pdf-genai -e SUNO_API_URL=http://host.docker.internal:3001/api/ -e OLLAMA_BASE_URL=https://host.docker.internal:11434/ -e OLLAMA_MODEL=llama3.1:8b -e VISUAL_MODEL=minicpm-v -e CORBADO_PROJECTID=xxxx -e CORBADO_API_KEY=xxxx -e APP_PORT=3000 kenken64/chatgpt

docker run -p 3000:3000 -d --name kenken64-chatgpt -e PINECONE_API_KEY=xxx -e PINECONE_INDEX_NAME=pdf-genai -e SUNO_API_URL=http://host.docker.internal:3001/api/ -e OLLAMA_BASE_URL=http://host.docker.internal:11434 -e OLLAMA_MODEL=llama3.1:8b -e VISUAL_MODEL=minicpm-v -e CORBADO_PROJECTID=pro-0317338422706138772 -e CORBADO_API_KEY=xxxx -e APP_PORT=3000 kenken64/chatgpt

docker run --network host -d --name kenken64-chatgpt -e PINECONE_API_KEY=xxxx -e PINECONE_INDEX_NAME=pdf-genai -e SUNO_API_URL=http://host.docker.internal:3001/api/ -e OLLAMA_BASE_URL=http://host.docker.internal:11434 -e OLLAMA_MODEL=llama3.1:8b -e VISUAL_MODEL=minicpm-v -e CORBADO_PROJECTID=xxxx -e CORBADO_API_KEY=xxx -e APP_PORT=3000 kenken64/chatgpt

docker compose up -d --build

ENV PINECONE_API_KEY=${PINECONE_API_KEY}
ENV PINECONE_INDEX_NAME=${PINECONE_INDEX_NAME}
ENV SUNO_API_URL=${SUNO_API_URL}
ENV OLLAMA_BASE_URL=${OLLAMA_BASE_URL}
ENV OLLAMA_MODEL=${OLLAMA_MODEL}
ENV VISUAL_MODEL=${VISUAL_MODEL}
ENV CORBADO_PROJECTID=${CORBADO_PROJECTID}
ENV CORBADO_API_KEY=${CORBADO_API_KEY}
ENV CORBADO_FRONTENDAPI=${CORBADO_FRONTENDAPI}
ENV CORBADO_BACKENDAPI=${CORBADO_BACKENDAPI}
ENV APP_PORT=${APP_PORT}
ENV OLLAMA_HOST=${OLLAMA_HOST}
SQLLITE_DATABASE_DIR: ${SQLLITE_DATABASE_DIR}