Skip to content

Latest commit

 

History

History
56 lines (42 loc) · 2.69 KB

README.md

File metadata and controls

56 lines (42 loc) · 2.69 KB

Streaming API Examples

This directory showcases how the Behavioral Signals Python SDK can be used to interact with the streaming API. The examples demonstrate how to send audio data in real-time and receive behavioral analysis results.

Before running the scripts, ensure you have set up your environment correctly (note: we use Astral's uv as the default package manager):

uv venv -p python3.10 venv
source venv/bin/activate

uv pip install behavioralsignals python-dotenv

For convenience, all of our examples read your API credentials from the environment variables CID and API_KEY. You can either set them in your shell:

export CID=your_cid
export API_KEY=your_api_key

or create a .env file in the same directory as the scripts with the following content, which will be automatically loaded by the examples:

CID=your_cid
API_KEY=your_api_key

🎤 Record from microphone

You can use the streaming_api_microphone.py script to capture audio directly from your microphone and stream it to the Behavioral Signals API in real time. This is a common use case when integrating with audio streaming APIs for live analysis or feedback.

First, make sure you install some additional dependencies required for audio capture and live display:

uv pip install sounddevice numpy rich

Then, run the script and start speaking into your microphone. The results will be displayed in a live table format, showing the analysis of your speech in real-time.

python streaming_from_mic.py --api behavioral --response_level all
  • The --api argument specifies which API to use (either behavioral or deepfakes).
  • The --response_level argument controls the level of response granularity. Options include segment, utterance and all:
    • segment: Provides segment-level results, i.e. interim results for 2-second segments of continuous speech
    • utterance: Provides utterance-level results, i.e. results for complete utterances. Conceptually, an utterance is made up of one or more segments (though with --response-level utterances, only the uttterance-level results are returned).
    • all: Provides all available results for each utterance, including segment-level results and utterance-level.

📁 Stream from file

Use streaming_api_file.py to send audio data from a file to the Behavioral Signals API in real-time. This has limited usability but we include it for completeness.

python streaming_api_file.py --file audio.wav --output results.json --api behavioral --response_level all

The results will be printed to the console in raw format as they are processed, and will be also saved to a file named results.json.