VibeCore OS // Autonomous AI Software Factory

An automated AI assistant that writes code, runs tests, and fixes programming errors automatically while keeping cloud costs low.

The Problem

Writing repetitive boilerplate code wastes developer hours, while continuously querying cloud AI tools (like OpenAI or Gemini) balloons API costs rapidly.

The Solution

VibeCore OS coordinates a local-cloud developer loop, executing quick code drafts using free local AI models, and only routing complex refactoring tasks to cloud APIs.

Local-Cloud Hybrid Architecture

Everyday code writing is handled completely offline. If the system fails a test or hits a logic block, it formats the error and queries cloud AI to generate a correction payload, reducing API billing rates.

3. Private Testing & Self-Healing

It creates a private workspace to write and test code. It runs checks in this hidden space first. If something fails, the AI automatically corrects the code. The finished code is only saved once all checks pass.

PROJECT METADATA

STATUS Active Utility

TYPE AI CLI Agent

TECH STACK Python CLI
Ollama Local API
Gemini Cloud API
Shadow Merging

CLI Sandbox Terminal

VibeCore OS active. Run "vc.py status" or "vc.py sprint" to begin.

Spent Limit: ₦10,000.00

Budget Caps: ₦10,000 Limit

Local Calls: 42

Cloud Calls: 7

The ₦10,000/Month Strategy

The AI tracks its own cost. If cloud API spending exceeds ₦10,000, it automatically turns off cloud tools and switches completely to free, local tools. This protects developers from unexpected bill surprises.

VibeCore LLM-Cloud Task Sandbox

INTERACTIVE DEPLOYMENT

Write a simple coding goal below (e.g., build a payment router or write tests) and click "Run objective". You will see the local AI draft the code, run checks, and call the cloud AI for help if it gets stuck.

Objective Prompt

Preset Prompts

Live Agent Trace Monitor AWAITING OBJECTIVE

[VibeCore Sandbox v1.2.0 active]
Objective input required. Select a template or write a custom goal, then hit RUN.