I’m here today to share the new feature from project named dochelper.ai .
What is it?
The dochelper.ai is a tool using ai to make simple final user search info on your website or documentation from your product. But it is not only good for the final user, it is simply for the company to add new info to the AI about new products, features or extra info.
What is the new feature?
The new feature allows the chatbot to collect messages the chatbot can’t answer, because it doesn't have info to answer the question, so the chatbot will save on the database a question without an answer.
You can access the list message without an answer to see and understand what the users are asking and add the info on chatbot base of knowledge, so after that the chatbot will enable you to answer the question.
What are use case?
Documentation from your product Chatbot agent to answer common question about the production on Website Chatbot agent to help the internal team from the company. PS: probability you know or you are a guy who has a specific knowledge about some process or wait to execute something, so you can transfer that knowledge to chatbot that way you don't need to bottle or lose time waiting for the person to answer some.
Technologies? FireCrawl(extract children links from the page and extract content from the page without problem like: recaptcha, ip blocker or something else) OpenAI(embedding model and chat model). Supabase + Postgresql(RAG Retrieval-augmented generation). N8N workflow where they play as agents to answer questions about the product or service.
Link of project: https://github.com/tiago123456789/dochelper.ai Link of Dochelper.ai v2: https://github.com/tiago123456789/dochelper.ai/blob/main/Dochelper___Simple_chatbot_support___AI___RAG___Firecrawl_v2.json Link of video explaining: https://youtu.be/CquRZPke8Y0