GitHub Repo: https://github.com/daniel-manickam/boxbeyond-kutiraai
KutiraAI is an open source local AI platform designed for privacy-first enterprise AI deployments. This cutting-edge AI platform runs entirely on-premises, combining advanced AI with Apple Silicon optimization (MLX-LM) for fast, secure fine-tuning of large language models (LLMs). KutiraAI features dataset generation, fine-tuning, CrewAI multi-agent workflows, n8n for business process automation and Qdrant vector search for Retrieval-Augmented Generation (RAG) — all containerized for enterprise scalability
Technical Stack:
React (Frontend), MLX-LM (Fine tuning), Celery (async – long running jobs), Dataset Generation (Local & Google Drive – N8N process flows), Agentic workflow (UI + CrewAI), Business Process (N8N), Qdrant (Vector Store), Redis (Cache), Docker, Ollama
Key Capabilities:
1. Dataset Generation: Automatically generate training datasets from your documents (Google Drive, local files)
2. Local Fine-tuning: Train LLMs on your proprietary data using Apple’s MLX-LM library.
3. Agentic Workflows: Create intelligent multi-agent workflows powered by CrewAI
4. Process Integration: Seamlessly integrate agentic workflows into enterprise processes with n8n workflows
5. Vector Store: Integrated Qdrant vector database for RAG and semantic search
For any business inquiries or partnership opportunities, please contact me at daniel@boxbeyond.ai
Source
