Open-source Linklbee Scan finds LLM API billing leaks from local logs

Linklbee Scan is a new open-source Python tool that analyzes existing LLM API usage logs on a user's machine to pinpoint sources of unnecessary spending. Unlike simply viewing a bill, it helps users investigate questions such as whether repeated requests could be reused, whether prompts hit the provider's cache, and whether failed or truncated calls are still being charged. The tool runs locally using the Python standard library, requires no API key or model calls, and produces an HTML report with aggregate figures rather than prompt or answer text. Each check is marked as found, none, or needs more data, and the scanner supports JSONL, CSV, and other common log formats from providers and tools like Langfuse and LiteLLM. Published under the MIT license by the NOTEGG team, the project invites user feedback on report usefulness and field mapping accuracy.