IJFW

•
161
•

Unifies and enhances the capabilities of multiple AI coding agents through shared memory, smart routing, and disciplined workflows to optimize development and reduce token costs.

IJFW transforms the experience of using AI coding agents by addressing common issues like forgetfulness, lack of discipline, and inefficient token usage. It provides a single install solution that enables six popular AI agents—including Claude Code, Codex, Gemini, and Copilot—to share a persistent local memory across sessions and projects. By implementing smart routing, disciplined workflows, and on-demand multi-AI audits, IJFW ensures that tasks are dispatched to the right model, outputs are concise, and blind spots are caught, ultimately reducing operational costs while significantly improving code quality and developer productivity.

Key Features

01Shared, Persistent Local Memory: Unifies knowledge across all AI agents and sessions.
02Disciplined Workflow Engine: Guides development with structured Quick (FRAME, WHY, SHAPE, STRESS, LOCK) and Deep (FRAME, RECON, HMW, DIVERGE, CONVERGE, LOCK) modes.
03Multi-AI Cross-Audits and Critiques: Leverages multiple AI models for on-demand code review and problem identification.
04Token Economy Optimization: Employs smart routing, output rules, and prompt-cache discipline to reduce token expenditure.
05Real-time Observability Dashboard: Provides a live window into AI sessions, costs, and cache efficiency via a local web interface.
060 GitHub stars

Use Cases

01Streamlining AI-assisted software development lifecycle from conception to deployment.
02Improving cost-efficiency and performance of multiple AI coding agents across projects.
03Enhancing code quality and project discipline through automated multi-AI reviews and structured planning.