Getting started

Your first transcript with Audio Analyser

From a fresh clone to a transcript, an analysis and a set of recommendations.

What do I need?

  • Python 3.10 or newer. CI tests on Python 3.10.
  • PortAudio, for PyAudio: portaudio19-dev on Debian and Ubuntu, as CI installs it. PyAudio publishes wheels for Windows only, so elsewhere pip builds it against PortAudio.
  • Azure resources: a Speech resource for transcription and text to speech, a Language resource for analysis, and a Translator resource if you translate.
  • A recommendation provider: an API key for OpenAI, Anthropic or Gemini, or a local Ollama server.

How do I install it?

git clone https://github.com/sebastienrousseau/audioanalyser.github.io.git
cd audioanalyser.github.io
python -m venv .venv && . .venv/bin/activate
pip install -r requirements.txt

requirements.txt pins every dependency with hashes. For the Anthropic or Gemini providers, also install their SDK: pip install anthropic or pip install google-genai.

How do I configure it?

Copy the example environment file and fill it in:

cp env.example .env

The main settings:

Variable What it is
INPUT_FOLDER Where recordings are written and audio is read
TRANSCRIPTS_FOLDER Where transcripts are written
REPORTS_FOLDER Where text analysis results go
RECOMMENDATIONS_FOLDER Where summaries and recommendations go
AZURE_AUDIO_TEXT_KEY, REGION Azure AI Speech key and region
AZURE_LANGUAGE_ENDPOINT, AZURE_LANGUAGE_KEY Azure AI Language
LLM_PROVIDER, LLM_MODEL Recommendation provider and model

The reference lists the rest.

How do I run a first analysis?

python -m audioanalyser -rec            # record ten seconds
python -m audioanalyser -stt meeting.wav
python -m audioanalyser -ta
python -m audioanalyser -sum

Or start the dashboard and do the same from the browser:

python -m audioanalyser -s

then open http://127.0.0.1:8080/.