transcriber-sh 0.1.0

Real-time microphone → OpenAI → text (CLI)

# Transcriber

> Real-time **microphone + system-audio → OpenAI → transcript → summary** — all from a single CLI command.

---

* **One-step setup**  
  ```bash
  transcriber setup                     # prompts for API key & model – writes .env

• Automatic device routing – Captures the default microphone and the loop-back that matches your current output device (no index hunting). – Works on Windows, macOS, Linux (Pulse/ALSA).

• Realtime transcription (OpenAI Realtime API) – Streams 24 kHz PCM16 chunks; prints each finalized text fragment in the terminal.

• Instant summary – On Ctrl-C it saves a timestamped .txt transcript and generates a one-paragraph summary via OpenAI Chat.

• Pure Python, no drivers – uses soundcard + sounddevice for audio I/O.

---

🔧 Quick-start

# 1. install
pip install transcriber          # or: pip install git+https://github.com/Dogmeat0/transcriber.git

# 2. initial configuration
transcriber setup                # enter your OpenAI API key and choose model

# 3. record!
transcriber                      # press Ctrl-C to stop & get summary

Command	What it does
transcriber list-dev	Lists all microphones / loop-backs & the current defaults.

OPENAI_API_KEY = sk-************************************
MODEL          = gpt-4o-mini-transcribe


```bash
git clone https://github.com/Dogmeat0/transcriber.git
cd transcriber
python -m venv env && source env/bin/activate
pip install -e ".[dev]"             # editable + dev tools
ruff check transcriber              # lint
pytest                               # tests (if/when added)