gremlord

gremlord provides a flexible local router for running Claude Code with various underlying models. Key features include: * Compatibility with Anthropic, OpenAI, xAI, and Ollama endpoints * TUI and tool support identical to Claude Code * Precise token metering and budget enforcement * Session naming and inter-session communication * Open-source under an MIT license This utility acts as a thin wrapper, allowing users to leverage their preferred large language models while maintaining the familiar Claude Code interface and functionality. It handles request/stream translation for non-Anthropic endpoints and ensures byte-faithful passthrough for Anthropic models. Every token consumed is tracked and priced, with hard budget limits to prevent unexpected spending, giving users full cost control. gremlord is ideal for developers, researchers, and teams who want to standardize their coding environment around Claude Code's capabilities but require the flexibility to switch between different model providers or integrate local models. It simplifies testing, development, and deployment across diverse model ecosystems while managing operational costs effectively.