Leonxlnx/earshot ? reverse-engineered prompt

Reverse engineered prompt

Build me a local Python tool called earshot that helps coding agents understand sound files and make better audio decisions.

I want it to run on CPU with no network needed for the core features, and expose a command line, a Python API, and an MCP server so I can use it from an agent. The main thing should be to look at an audio file and explain it in plain language, rank how well it matches a description, find the best moments in a longer recording, compare two sounds, and tell me if something feels like a real recording or something made in code. It should also be able to turn a recording into a code recipe that can be edited and rendered again, and include a small built in library of sound effects that can be rendered and checked.

Please make the output friendly and easy to read, with JSON available too if needed. If you need to look up current docs or best practices, go ahead and do that.

Are you gonna build this?

make sure you review the code using coderabbit

Try freeSponsored — opens CodeRabbit in a new tab