Quick Start

1. Configure

Copy .env.example and set your LLM API key:

cp .env.example .env

Minimum required:

AUDIA_LLM_PROVIDER=openai
AUDIA_OPENAI_API_KEY=sk-...

2. Convert a local PDF

audia convert paper.pdf

Output: ~/.audia/audio/<stem>_<timestamp>.mp3

3. Search ArXiv and convert

audia research "attention is all you need" --max-results 3 --convert

4. Voice input

audia listen

Record your query (up to 30 s), confirm the LLM-distilled search terms, pick papers, convert.

5. Web UI

audia serve
# → http://127.0.0.1:8000

6. Show active config

audia info

Output location

All audio files land in ~/.audia/audio/. Debug snapshots (raw / preprocessed / curated text) are saved under ~/.audia/debug/<run_id>/. Override the root with AUDIA_DATA_DIR=/your/path.