Turn your Google Drive files into a fast AI chat that anyone on your team can use. The flow collects documents from a chosen folder, enriches them with helpful tags, and saves them in a vector database so answers come back quickly and with context.

The system starts from a manual button, a webhook call, or a chat message. It looks up file IDs in Google Drive, downloads content, and extracts key details like main themes and topics. Text is split into smart chunks and turned into embeddings with OpenAI. Those chunks and metadata are stored in Qdrant for hybrid search. A chat agent then uses Qdrant to find the right passages and forms a clear reply using OpenAI or Google Gemini. Chat history is saved to Google Docs. A Telegram approval step protects your data when deleting items from the vector store.

You only need to set your Google folder ID and the Qdrant collection name, connect your accounts, and run a test. Most teams see faster answers, less time hunting for files, and safer data updates. It fits knowledge search, policy lookup, onboarding guides, and draft support replies. Once live, it scales to many files and users with minimal upkeep.

• Multi entry triggers: run by manual click, webhook, or incoming chat message
• Google Drive folder scan to find file IDs, then download and read file content
• Automatic metadata extraction to capture themes and recurring topics
• Token based text splitting and OpenAI embeddings for precise vector search
• Qdrant insertion with collection control to build a scalable knowledge index
• Chat agent with memory that queries Qdrant and responds using OpenAI or Google Gemini
• Google Docs logging to keep a running chat history you can review
• Telegram confirmation to approve or decline deletion of Qdrant points

• Reduce manual document search from hours to minutes
• Streamline knowledge lookup by up to 70 percent with chat retrieval
• Improve answer accuracy by enriching each chunk with extracted metadata
• Handle 10 times more documents with batch ingestion and vector search
• Connect Google Drive, Qdrant, OpenAI, Google Gemini, Google Docs, and Telegram in one flow
• Lower risk with human approval before deleting data from the vector store

1. Import the template into n8n: Create a new workflow in n8n > Click the three dots menu > Select 'Import from File' > Choose the downloaded JSON file.
2. You'll need accounts with Google Drive, Google Docs, OpenAI, Google Gemini, Qdrant and Telegram. See the Tools Required section above for links to create accounts with these services.
3. In the n8n credentials manager, open any Google Drive node, click the Credential to connect with dropdown, select Create new credential, and follow the on screen steps to grant access to your Drive. Use a workspace account with read access to the target folder.
4. Open the Google Docs node and create a new credential the same way. Allow the scopes needed to create and edit the chat log document.
5. For OpenAI, create an API key from your OpenAI account, then in any OpenAI node choose Create new credential and paste the key. Name the credential clearly, for example openai main.
6. For Google Gemini, get an API key from Google AI Studio. In the Gemini nodes, create a new credential and add the key. Pick the model configured in the node if you want to match defaults.
7. For Qdrant, collect your API URL and key from your Qdrant Cloud project or your self hosted instance. In the Qdrant nodes, create a new QdrantApi credential and enter the URL and key. If you use the delete code node, open it and update the URL and key values inside the code to match your environment.
8. For Telegram, create a bot using BotFather and copy the bot token. In each Telegram node, create a new telegramApi credential with that token. Start a chat with your bot and capture the chat ID if the nodes require it.
9. Set your folder and collection: open the Set node named Google Folder ID and paste your folder ID. Open the Set node named Qdrant Collection Name and enter your collection name.
10. Check data handling settings: in the Token Splitter node confirm chunk size is set to 3000. In the embeddings node confirm the model matches your plan. In the Extract Meta Data node review the attributes list.
11. Run a test ingestion: click Test workflow, watch the Google Drive files load, and confirm new points appear in your Qdrant collection. If nothing loads, verify folder ID and permissions.
12. Test the chat: trigger the chat input and ask a question covered by your files. Confirm a clear answer and see a new entry in Google Docs for the chat history.
13. Test safe deletion: build a file ID list, run the delete path, then approve or decline in Telegram. Verify the related points are removed from the Qdrant collection when approved.
14. Troubleshoot common issues: 401 errors mean invalid API keys, 404 on Qdrant means wrong collection name, empty answers suggest missing embeddings or chunking too large. Adjust settings and retest.