An AI-driven LINE chatbot providing different news perspectives: AI, what's your take?
Have you ever read a news article and felt that something about the report was a bit off, but couldn't quite put your finger on it?
If you're interested in hearing alternative viewpoints, this LINE chatbot is more than happy to offer its insights 🎙️.
Simply share a news link with it, and it will provide comments based on the article's headline, perspectives, and presentation, offering them as a reference for you.
Of course, it's important not to fully accept its opinions, as AI can easily make mistakes, and its comments might not always be impartial 😅. This isn't professional news critique; the AI's viewpoints are merely for consideration, so please don't take them too seriously
Additionally, because the bot is powered by the same GPT-3.5 as ChatGPT 🧠, feel free to have casual conversations with it or ask questions beyond news!
Click to add as a friend👉👉 AI, what's your take?
- CHANNEL_SECRET The key for the LINE Messaging API, which can be obtained as outlined in Connecting ChatGPT to LINE.
- CHANNEL_ACCESS_TOKEN Another key for the LINE Messaging API, obtainable following the instructions in Connecting ChatGPT to LINE.
- OPENAI_API_KEY The OpenAI API key, attainable as explained in Connecting ChatGPT to LINE.
- MONGO_CONNECTION_STR The connection key for the MongoDB database, details available at An Introduction to MongoDB Connection Strings.
- TEMPERATURE The response temperature for the OpenAI API, ranging from 0 to 2, as described in the API reference - OpenAI API.
- MAX_TOKEN_LIMIT The number of tokens to retain in the conversation memory, consult LangChain- ConversationTokenBufferMemory.
pip install -r requirements.txt gunicorn app:app --timeout 3600(Due to occasionally long response times from the OpenAI API, a longer timeout is set.)
- Repositories
- Articles/Documentation
Pull requests are welcome! For significant changes, please open an issue to discuss the proposed modifications.
This project's flexibility and scalability are substantial. Changing a prompt immediately transforms it into a different type of bot, and the use of LangChain framework makes adding advanced features like embedding queries a straightforward task.
Feel free to use the code from this project to create various useful conversational bots by unleashing your imagination.