This project provides a unified API proxy for various AI challenges in text, audio, image, and video understanding and generation, aiming to consolidate well-known solutions for easy client access. It's designed for developers and researchers seeking a consolidated interface to AI functionalities, with a long-term goal of enabling offline, on-device AI capabilities, particularly for the SUSI project.

How It Works

The API is structured around challenge fields (text, audio, image, video) and mirrors existing provider API definitions for consistency. It includes drop-in replacements for specific OpenAI API endpoints, facilitating a smooth transition. The project's ultimate aim is to replace these proxy functions with self-hosted AI models, leveraging recent advancements in transformer models for feasibility.

Quick Start & Requirements

• Install dependencies: pip3 install -r requirements.txt
• Start the API server: python3 src/main.py --openai_api_key <OPENAI-API-KEY>
• Requires an OpenAI API key for initial use. Future versions will offer open replacements.
• Docker image can be built with docker build -t susi_api . and run with docker run -d -p 8080:8080 -e OPENAI_API_KEY=<apikey> --name susi_api susi_api.
• Testing involves cd test/src && ./test_voice.sh.

Highlighted Details

• Aims to replace proxy functions with on-site hosted AI functions.
• Re-implements specific OpenAI API endpoints for compatibility.
• Docker image includes pre-loaded voice decoder models (tiny and base).
• Long-term goal of complete offline functionality on devices like Raspberry Pi.

Maintenance & Community

No specific contributors, sponsorships, or community links (Discord/Slack, roadmap) are mentioned in the README.

Licensing & Compatibility

The README does not explicitly state a license. Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

The project currently requires an OpenAI API key, though future versions plan to offer open replacements. The README does not detail specific AI models used or performance benchmarks.