GPT Interaction and Application Platform

Project Overview

This is a Flask-based web application that provides API endpoints and a simple front-end interface for interacting with GPT models. Users can chat with AI, generate images, or get content summaries via a WeChat official account or web interface.

Key Features

• Text conversation with GPT models: Supports multi-turn conversations for human-like chat experiences.
• Image generation from user input: Generate high-quality images using AI models.
• WeChat official account message handling: Supports auto-replies, image generation, and other features.
• Content summarization: Extract key information from long texts or web pages and generate concise summaries.
• Custom AI service providers: Supports custom AI providers, including OpenAI-compatible APIs.

One-Click Deployment

Note: You need to configure the environment variables according to your actual usage.

Usage

Environment Setup

1. Clone the repository:

   git clone https://github.com/ryanuo/gpt.git
   cd gpt

2. Install dependencies:

   pip install -r requirements.txt

3. Configure environment variables:

   cp .env.example .env

4. Start the service:

   python -m api.index

Environment Variables

Variable Name	Description	Example Value	Required	Notes
WX_TOKEN	WeChat official account token, used for message signature verification	wechat_token_123456	Yes (for WeChat)	Only needed if using WeChat message handling
OPENAI_API_KEY	API key for GPT models	sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx	No	Required when using a custom OpenAI-compatible API (PROVIDER=custom)
PROVIDER	Specify AI service provider type	custom	No	Default is empty (uses g4f client default). custom means using a custom API
COMPLETION_MODEL	Default text generation model	gpt-4o-mini, gpt-3.5-turbo	No	Defaults to default_model in api/config.py (gpt-4o-mini)
API_KEY	Authentication key for custom AI service	custom_api_key_789	No	Required if PROVIDER=custom, similar to OPENAI_API_KEY
API_URL	API endpoint URL for custom AI service (OpenAI-compatible)	https://api.example.com/v1	No	Required if PROVIDER=custom, defaults to OpenAI official API

Additional Notes

• Environment variables should be configured in the project root .env file (referenced in .gitignore to prevent leaks).

• Example setups:

  • Using default g4f service: no need to set PROVIDER, API_KEY, API_URL; optionally set COMPLETION_MODEL.
  • Using a custom API: set PROVIDER=custom, API_KEY, API_URL; optionally set COMPLETION_MODEL.

API Endpoints

API Type	Path	Method	Parameters	Response
Conversation	/g4f/<model>	POST	- message: user input text - context (optional): conversation context (format [{"role":"user","content":"xxx"},{"role":"assistant","content":"xxx"}]) Note: Model specified via <model> path; available models via /vmodels	JSON containing response field (model-generated reply)
Image Generation	/generate-image	POST	- prompt: description of the image (e.g., "Cyberpunk city night scene") Note: uses flux model	JSON containing image_url field (URL to access generated image)
WeChat Official Account	/wechat	POST	Automatically parses WeChat server push messages (includes signature, timestamp, nonce, XML content, etc.) Note: uses default model in api/config.py, gpt-4o-mini by default	Returns XML message reply according to WeChat specifications (text or rich media with image URLs)
Content Summarization	/ai-post	POST	- url: URL of webpage to summarize (e.g., https://example.com/article) Note: uses default model in api/config.py, gpt-4o-mini by default	JSON containing summary field (concise summary of key content, ~50-200 words)
Model List	/vmodels	GET	None Note: returns only supported model list, no model invocation	JSON grouped by type, e.g.: { "GPT Series": ["gpt-4o-mini", "gpt-3.5-turbo"], "Other Models": ["claude-3-haiku", "gemini-pro"] }

References

• wechatpy Documentation
• OpenAI GPT Models

License

This project is licensed under the MIT License. See LICENSE for details.