AI is becoming part of everyday life, but many people still cannot afford AI subscriptions or continuous API usage.
I started building Mrigashira AI with a simple idea:
AI should be accessible to everyone, not just those who can afford a subscription.
Mrigashira AI is a desktop AI engine designed to run LLMs locally on your own hardware, while also giving you the option to connect to cloud AI providers when needed.
With local models, you can:
It also includes a Local Model Manager for down and managing local models.
One of the main challenges I focused on was running AI on ordinary laptops.
Mrigashira AI can dynamically adjust the context window based on available RAM/VRAM, helping make local inference more practical on systems with limited resources.
When local AI isn't enough, you can connect your own API keys to providers such as:
OpenAI • OpenRouter • Ollama Cloud • Grok
This gives you the flexibility to choose the right model for the task.
Mrigashira AI also includes:
The Office plugin supports PDF, DOCX, PPTX, XLSX, ODT, ODS, ODP and CSV.
The plugin architecture opens possibilities for:
Local RAG • Code Assistants • Agent Workflows • Office Automation • Custom AI Tools
My long-term goal is to build a foundation for a local AI ecosystem that is more accessible, private, and less dependent on centralized infrastructure.
🇮🇳 Released on India's 80th Independence Day.
🌐 **Try it:** [https://mrigashira.co.in](https://mrigashira.co.in)
🎥 **Presentation:** [https://youtu.be/mQywg3Daey8](https://youtu.be/mQywg3Daey8)
📖 **Installation Guide:** [https://youtu.be/oClOTeMPpko](https://youtu.be/oClOTeMPpko)
I'd love to hear what developers think and what you'd like to build with local AI.
#AI #LocalAI #OfflineAI #LLM #GenerativeAI #Privacy #DevTools #AIForAll #MadeInIndia