🧠 Store unlimited knowledge.❓When a user asks a question, 🔑 access the relevant knowledge and supply it to 🤖 ChatGPT so it can create a human-like answer 💬 (based on that knowledge context).
〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰

This plugin uses & requires 2 services:

1️⃣ Pinecone.io
2️⃣ OpenAI

〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰

The flow looks something like this:

🗄 Vectorize your knowledge & store it in the Pinecone database

(using the “Pinecone - Upsert” action step)

🙋‍♂️ When a user asks a question, convert that question into vector embeddings (meaning, but formatted as a bunch of numbers)

(using the “OpenAI - Get Embeddings” action step)

🧐 Pass those numbers to Pinecone, so it can grab the top bits of knowledge that are most closely related to your user’s question.

(using the “Pinecone - Query” action step)

🚀 With pinecone’s response, you can now supply those most relevant bits of knowledge (as context) back over to ChatGPT, along with the user’s prompt/question, so ChatGPT can respond with accurate information in a human-like manner.

(using the “OpenAI - Get Answer” action step)

〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰

Yes, I'm aware that there are like 700 api-key inputs down here 👇 for API keys (sorry, it's necessary to have both services in 1 plugin).

You really only need 2 keys, though: 1 from Pinecone and 1 from OpenAI.

The Pinecone key can be pasted directly into each pinecone field.

The OpenAI key requires "Bearer " to precede it. So be sure to type Bearer (then a space) and then paste the key. Do that for each OpenAI input down below. 👇

〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰〰

If you get stuck or want to add a feature or need help, we'd love to have you join our growing academy at NoCodeAdvantage.com!