Welcome to the Peridot Kernel

[uncoalesced]

SOVEREIGN AI KERNEL | PERIDOT

Welcome to the Peridot communications node. This repository is dedicated to building a 100% local, air gapped LLM runtime that leverages idle GPU compute for medical research.

> PROTOCOLS FOR ENGAGEMENT

1. No Cloud APIs: Peridot is strictly local. Do not propose integrations with OpenAI, Anthropic, or external telemetry services.
2. Hardware Specifics: When reporting performance, always list your OS, CPU, GPU, and VRAM.
3. Medical Research: If you are running the Folding@home module, share your stats in the [RESEARCH] category.

> CURRENT ARCHITECTURE STATUS

• Engine: Llama-3 (8B) CPU-bound / GPU-offloaded
• Audio: Localized Whisper via FFmpeg
• Vision: Offline (Severed for VRAM stability)

Search the [SUPPORT] Q&A before opening a new thread.

Engineered by uncoalesced.