This project provides a compatibility layer to expose Pandora-Next's backend API as a standard OpenAI v1 API, specifically targeting /v1/chat/completions and /v1/images/generations. It's designed for users who want to integrate Pandora-Next with other applications that expect OpenAI's API format, offering features like streaming, DALL-E image generation, and support for various GPT models and GPTS.

How It Works

The core functionality involves proxying requests from the standard OpenAI v1 endpoints to Pandora-Next's internal /backend-api endpoints. It handles both streaming and non-streaming responses for chat completions and supports DALL-E image generation requests. The project also includes features for API key management for GPTS, custom endpoint prefixes, and a "Bot Mode" for integrating with chat bots, which allows for customized output formatting of tool usage and results.

Quick Start & Requirements

• Install/Run: Deploy via Docker Compose using the provided docker-compose.yml.
• Prerequisites: Requires a running Pandora-Next instance with auto_conv_arkose:true enabled.
• Configuration: Modify environment variables in docker-compose.yml for Pandora-Next URL, API prefix, logging, worker/thread counts, and model aliases.
• Docs: Release Page, Community Telegram Channel

Highlighted Details

• Supports various models including gpt-4-s, gpt-4-mobile, gpt-3.5-turbo, and GPTS.
• Enables DALL-E image generation and file/URL input for vision models.
• Offers a "Bot Mode" for enhanced integration with chat bots, controlling output formats for tools and images.
• Includes options for custom Arkose token URLs and using Pandora-Next's file server.

Maintenance & Community

The project acknowledges contributions from various individuals. A Telegram channel is available for community discussion and support.

Licensing & Compatibility

The project is released under the MIT License, allowing for commercial use and integration with closed-source applications.

Limitations & Caveats

The project explicitly states it does not bypass OpenAI or PandoraNext official limitations and is not designed for high-concurrency scenarios due to its proxy nature. It also warns that using generated Arkose tokens may lead to account bans, with all consequences borne by the user.