Turn website voice questions into clear answers from your own files. This build links a voice agent to a retrieval system so visitors get fast, accurate replies based on content in your Google Drive. It fits teams that want self service support without growing chat queues.

Here is how it runs. The website voice widget sends each question to an n8n webhook. An AI Agent uses OpenAI to read the question, fetches the most relevant passages from a Qdrant vector database, and forms a helpful answer. Your documents are loaded from Google Drive, split into small parts, turned into embeddings by OpenAI, and stored in Qdrant. The reply is sent back through a response webhook so the voice agent can speak it.

Setup is straightforward. You need access to Google Drive files, an OpenAI API key, and a Qdrant cluster and API key. You also set up an ElevenLabs agent and paste the n8n webhook URL into its tool. Expect fewer repetitive tickets, faster answers, and a better user experience. Great for FAQs, menus, policies, product docs, or onboarding guides presented as voice Q and A.

• Webhook listener receives the user question from the website voice widget
• AI Agent powered by OpenAI crafts the final response with short term memory
• Vector Store Tool searches Qdrant for the best matching content
• Google Drive integration fetches a folder and downloads files for indexing
• Document loader and token splitter prepare text into small chunks
• OpenAI embeddings convert chunks into vectors stored in Qdrant
• HTTP requests create or clear the Qdrant collection before indexing
• Respond to Webhook sends the answer back for voice playback

• Reduce manual FAQ handling by up to 80% with always on voice answers
• Cut first response time from minutes to seconds for common questions
• Handle many visitors at once without adding staff
• Keep answers consistent by pulling from a single source of truth in Google Drive
• Improve accuracy by retrieving the most relevant document passages
• Connect voice, AI, and vector search without custom code

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, OpenAI, Qdrant and ElevenLabs. See the Tools Required section above for links to create accounts with these services.
3. In n8n Credentials, connect Google Drive: open either Google Drive node > Credential to connect with > Create new credential > follow the on screen steps to sign in and grant access.
4. Create an OpenAI credential: open an OpenAI node > Create new credential > paste your API key from the OpenAI API keys page > save.
5. Set up Qdrant access: open the Qdrant Vector Store or HTTP Request nodes > Create new credential. For API key or header auth, paste your Qdrant URL and API key. Use a clear name like Qdrant Prod.
6. Configure collection values: replace QDRANTURL and COLLECTION in the HTTP Request nodes and set the same collection name in both Qdrant Vector Store nodes.
7. Choose the Google Drive folder to index: in Get folder select the folder, then confirm Download Files pulls the right files.
8. Build the index: click Test workflow to run Create collection, clear old points, load files, split text, create embeddings, and upsert into Qdrant. Check node logs for success.
9. Create your voice agent in ElevenLabs: add a tool that sends a POST request to the n8n Webhook Listen URL. Include a body field named question that contains the user question.
10. Connect the reply path: keep the Respond to Webhook node in the chain so the text answer returns to ElevenLabs for voice playback.
11. Test end to end: click Test AI agent in ElevenLabs and ask a question. Confirm n8n receives the question and returns a spoken answer pulled from your Drive files.
12. Troubleshoot: if no matches are found, verify the Qdrant collection name, embeddings credentials, and that files were downloaded. If no voice plays, confirm the Respond to Webhook node runs and returns 200.