CouchSurfer.co.uk
Travel · Encryption · Running · Programing · Spain · Cyber · Building in public

Tools · Built in public · 2026

Sound in,
meaning out.

A three-stage voice pipeline, each stage a standalone tool. Capture sound, turn it into a transcript, then hand it to an LLM. Use one stage or chain all three.

LLM
Sound Audio file Transcript LLM
01

Stage one · Capture

soundToAudio

Capture live sound from the microphone and turn it into a clean audio file, ready for the next stage.

In: microphone  ·  Out: audio file

Open tool →
02

Stage two · Transcribe

audioToTranscript

Take an audio file and run it through speech-to-text, producing a plain transcript you can read, edit, or pipe onward.

In: audio file  ·  Out: transcript

Open tool →
03

Stage three · Understand

transcriptToLLM

Feed the transcript to an LLM — summarise, extract actions, answer questions, or do something stranger.

In: transcript  ·  Out: whatever you ask for

Open tool →