I’m building SAGE GUI - Electron, a local-first desktop app for AI coding workflows.
SAGE lets you run commands safely with:
pip install psycgod-sage
Or
npm install -g psycgod-sage Or
Use MCP Server It tracks command output, compresses noisy logs, shows tokens used, estimates token savings/cost savings, and keeps raw data local.
The GUI will support multiple AI providers and keys. With API Traveller, users can add multiple API keys/base URLs, rotate automatically when one fails, and route across providers like OpenRouter, Gemini, Groq, and Ollama.
Inside the app, SAGE can show:
SAGE GUI is still in the making, and I’m aiming to publish it by August.
Repo: [https://github.com/PsYcGoD/sage](https://github.com/PsYcGoD/sage)
Live Public dashboard: [https://sage.api.marketingstudios.in/](https://sage.api.marketingstudios.in/)
Glama MCP: [https://glama.ai/mcp/servers/PsYcGoD/sage](https://glama.ai/mcp/servers/PsYcGoD/sage)
SAGE is for developers who want AI coding agents to be more local, trackable and less messy.
source & further reading
dev.to — original article
Nobody Taught You How to Use This Thing
Your AI agent checked its queue, found nothing, and went back to sleep. The queue was full.
I built a production AI agent for our NOC ticket queue in one shift