L2M: Claude Code but for legacy code modernization (github.com)

Legacy2Modern (L2M) is an open-source AI coding agent that helps you modernize legacy codebases from the terminal. It provides a clean TUI with interactive chat so you can ask questions, request refactors or translations to modern languages, and get streaming, markdown-rendered responses in real time. L2M integrates with your repo—automatic commits, undo support and repository-aware context—lets you add/remove files to the active session, and keeps persistent session history across runs. It’s Python 3.10+ and distributed under Apache-2.0; install via curl | bash or pip, and configure BYOK API keys in a .env (supports OpenAI, Anthropic, DeepSeek, Gemini and 100+ providers via LiteLLM). For AI/ML engineers and developer teams the significance is practical and strategic: L2M accelerates large-scale legacy modernization by combining conversational code understanding with repo-aware, versioned edits—reducing manual refactor risk and audit friction. Multi-provider support avoids vendor lock-in and lets teams choose models for cost, latency, or privacy; BYOK and local provider adapters can help meet compliance requirements. Streaming responses, file-scoped context, and git integration make iterative, reviewable changes feasible in CI/CD workflows. As an open project, it’s extensible for new languages, model backends, or enterprise integrations.