What is Audio Analyser?
A Python application that records, transcribes and analyses audio with Azure AI, and writes recommendations with a language model. See the features.
Do I need Azure?
Yes, for transcription, analysis, translation and text to speech. Only the recommendations step can run without a cloud service, using Ollama.
Can I install it from PyPI?
PyPI has version 0.0.6 from February 2024, without the provider choice. Install from a clone; see getting started.
Can transcripts stay on my machine?
The recommendations can, with LLM_PROVIDER=ollama. Transcription itself always goes through Azure AI Speech. See the security page.
Which language model does it use?
OpenAI's gpt-4.1-mini by default; Anthropic, Gemini, Ollama and two CLI providers are also supported. See providers.
Is there a graphical interface?
Yes: python -m audioanalyser -s starts a local dashboard on port 8080. See the dashboard.
Which Python versions are supported?
Python 3.10 or newer; CI runs on 3.10.
What licence is it under?
The repository carries the Apache License 2.0 and the MIT licence. See the terms.