🔧 Deployable On-Premises RAG

🕛 Zeit seit Veröffentlichung: 49 Tage, 15 Stunden 29 Minuten
📆 Veröffentlicht am: 04.03.2025 um 08:19 Uhr
Nachrichtenbereich: 🔧 Programmierung
🔗 Quelle: dev.to

I’m excited to introduce Minima, an open-source Retrieval-Augmented Generation (RAG) solution designed to work seamlessly on-premises or with integrations like ChatGPT and the Model Context Protocol (MCP). Whether you’re looking for a fully local RAG setup or prefer to integrate with external LLMs, Minima has you covered.

Minima is a containerized RAG solution that prioritizes security, flexibility, and simplicity. You can run it fully locally or integrate it with external AI services, depending on your needs.

Key Features

Minima currently supports three modes of operation:

Isolated Installation

• Fully on-premises operation with no external dependencies (e.g., ChatGPT or Claude).

• All neural networks—LLM, reranker, and embedding—run on your cloud or local PC.

• Ensures your data stays secure and private.

Custom GPT

• Query your local documents directly through the ChatGPT app or web interface via custom GPTs.

• The indexer runs on your local PC or cloud, while ChatGPT serves as the primary LLM.

Anthropic Claude

• Use the Claude app to query your local documents.

• The indexer operates on your local PC, with Anthropic Claude as the primary LLM.

With Minima, you can enjoy a flexible RAG solution that adapts to your infrastructure and security preferences.

Would love to hear your feedback, thoughts, or ideas! Check it out, and let me know what you think.

Cheers!

https://github.com/dmayboroda/minima

...