What it does
Two n8n workflows. The first imports documentation from Google Docs, splits it, embeds it with OpenAI and stores it in a MongoDB Atlas vector store. The second answers Telegram messages using that documentation, then asks the user for approval or disapproval and saves the feedback with the conversation.
Use cases
- 01Answer staff questions from product documentation
- 02Index Google Docs into a searchable vector store
- 03Collect thumbs up or down on each bot answer
Connects
Telegram · Google Docs · Code · AI Agent · Embeddings OpenAI · OpenAI Chat Model · Recursive Character Text Splitter · Default Data Loader · MongoDB Chat Memory · MongoDB Atlas Vector Store