Scorpia Personal AI Coding Desk
A developer assistant workbench combining local offline GGUF models with cloud LLMs, equipped with terminal execution, workspace file tools, git operations, and push-to-talk speech input.
Overview & Introduction
Scorpia Personal AI is a developer workbench that integrates coding assistance directly with local workspace tools. It allows an operator to inspect file trees, review code diffs, run terminal commands, and apply patches using either fully offline local models (via GGUF / llama-server) or high-capacity cloud models (Gemini / Grok) from a unified browser desk.
The Operational Problem
- Pasting entire proprietary repositories or customer data into external public web chats risks data leakage.
- Standard chat windows have no awareness of directory structure, requiring manual copy-pasting of file after file.
- Developers need tools that can actually inspect git status, read exact file lines, and apply precise code diffs without leaving the editor.
System Architecture & Features
- Tetsuo Tool Loop: Seamless agentic execution allowing the assistant to list directories, grep files, view lines, and run bash tests safely.
- Hybrid Inference Engine: Toggle seamlessly between an offline local Jarvis model (running on-device when confidentiality is mandatory) and cloud LLMs when complex multi-file reasoning is required.
- Push-to-Talk Voice Input: Integrated speech recognition in the composer bar allows hands-free prompt dictation while reviewing code on screen.