On a Telegram message, an OpenAI agent answers from a MongoDB vector store of Google Docs and replies via Telegram.
On manual run, indexes documents into Pinecone, then a Gemini agent retrieves context and writes outputs to Google Docs, Sheets, and Drive.
On manual trigger, answers queries via a Gemini and OpenAI agent over a Pinecone vector store built from Google Docs, Sheets, and Drive.
On manual run, embeds documents into Qdrant, then generates content with Gemini and OpenAI and writes results to Google Docs.