OpenAI Chat App

A simple command-line chat application using the OpenAI API.

How to scan for AI-BOM

Use the Snyk CLI with an AI-BOM command:

snyk aibom --experimental --json

Pipe the Snyk CLI output of AI-BOM findings to jq to easily find and match your results:

snyk aibom --experimental --json  | jq '.components[] | select(."bom-ref" | startswith("model:")) | ."bom-ref"'

Get a fancy visual view of the AI-BOM findings using ai-bom-visualizer npm package:

snyk aibom --experimental --json  | npx ai-bom-visualizer --open

Prerequisites

• uv - Python package and project manager
  • Install via curl:

    curl -sSL https://install.astral.sh | python3 -

• An OpenAI API key

Setup

1. Clone the repository and navigate to the project directory

2. Set your OpenAI API key as an environment variable:

   export OPENAI_API_KEY="your_api_key_here"

3. Install dependencies:

   uv sync

Running the App

OpenAI Chat

uv run main.py

Type your messages and press Enter to chat with the AI. Type quit to exit.

Whisper Speech-to-Text

uv run whisper.py

This script demonstrates speech-to-text transcription and translation using OpenAI's Whisper model:

• Loads a fine-tuned Whisper model (whisper-medium-fleurs-lang-id) for language identification
• Streams French audio samples from the Common Voice dataset
• Translates the French speech into English text
• Uses the Hugging Face transformers and datasets libraries

Note: The first run will download the model (~1.5GB) and stream audio samples.

BLIP Visual Question Answering

uv run blip.py

This script demonstrates visual question answering (VQA) using Salesforce's BLIP model:

• Loads the BLIP VQA model (blip-vqa-base) from Hugging Face
• Downloads a sample image from the web
• Asks a question about the image ("how many men are in the picture?")
• Returns a natural language answer based on the image content

Note: The first run will download the model (~1GB).

Examples

Other example projects you can clone and scan for AI-BOM findings:

• https://github.com/snyk-labs/ai-buffet/
• https://github.com/aws-samples/generative-ai-on-amazon-sagemaker/
• https://github.com/aws-samples/amazon-bedrock-workshop