Scribe

Route bulk file operations and documentation generation from Claude to cheaper OpenAI-compatible LLMs, reducing token costs by 60–90% on routine tasks.

Scribe acts as a cost optimization proxy, intercepting Claude's bulk file reads, documentation generation, and boilerplate writing requests and routing them to cheaper OpenAI-compatible models. Three tools handle bulk file reading and summarization, documentation writing, and boilerplate code generation. Benchmarks show 41% fewer output tokens and 13% lower total cost compared to Claude-only workflows.